0001
 1                    UNITED STATES OF AMERICA
 2                     DEPARTMENT OF COMMERCE
 3                               AND
 4                THE INTERNET EDUCATION FOUNDATION
 5                              - - -
 6   
 7                   ONLINE PRIVACY TECHNOLOGIES
 8                  WORKSHOP AND TECHNOLOGY FAIR
 9                              - - -
10   
11                                Auditorium
12                                Department of Commerce
13                                  Building
14                                1401 Constitution Ave., N.W.
15                                Washington, D.C.
16                                Tuesday, September 19, 2000 
17   
18             The workshop was convened, pursuant to notice,
19   at 9:18 a.m.
20
21
22
23
24
25
0002
 1         P R O C E E D I N G S
 2                                       (9:18 a.m.)
 3                INTRODUCTION - GREGORY L. ROHDE,
 4       ASSISTANT SECRETARY OF COMMERCE FOR COMMUNICATIONS
 5              AND INFORMATION, ADMINISTRATOR, NTIA
 6             MR. ROHDE:  Good morning and welcome.  My name
 7   is Greg Rohde.  I'm  the Assistant Secretary of Commerce
 8   for Communications and Information, as well as the
 9   Administrator of the National Telecommunications and
10   Information Administration.  I am very pleased to welcome
11   all of you here this morning at this Online Privacy
12   Technology Workshop and Technology Fair.
13             Central to the mission of NTIA is to promote
14   electronic commerce, as well as to promote the development
15   of new technologies.  I am very pleased today that you can
16   all come here to join us as we carry out our mission to
17   look at how we can promote electronic commerce and develop
18   new technologies.
19             The purpose of today's workshop is to focus on a
20   particular area of electronic commerce and that is an area
21   that is so vitally important to a lot of consumers, and
22   that is privacy.  You don't need to look at national polls
23   to tell you that consumers care deeply about protecting
24   their private information on line.  You can simply look in
25   the comic strips.  When it reaches a comic strip such as
0003
 1   "Cathy," when she calls it "the information superhighway,"
 2   you know we've got an issue that is really close to the
 3   pulse of the American people.
 4             So I'm very pleased to today that we can be here
 5   to discuss this issue.  We have a number of panel
 6   discussions we're going to have throughout the day. 
 7   They're going to explore various aspects of this issue. 
 8   We also are very fortunate to have a Technology Fair which
 9   is occurring out in the foyer area of the building, and
10   this has been co-sponsored by the Internet Education
11   Foundation.  We're very pleased at their participation in
12   this event.
13             Without any further delay, I would like to
14   introduce the new Secretary of Commerce.  We're very
15   pleased that the new Secretary has chosen to join us here
16   for a few minutes and is going to give a keynote address. 
17   As many of you know, shortly, this summer, President
18   Clinton nominated Mr. Norman Mineta to succeed William
19   Daley as the Secretary of Commerce.  Mr. Mineta comes to
20   us from a long career on Capitol Hill.  He also began his
21   career as the Mayor of San Jose, and I can tell you that
22   since coming on board in the Commerce Department he has
23   certainly energized and inspired many here, and we're
24   extremely honored and pleased that Secretary Mineta is
25   here with us today.
0004
 1             So without any further ado, I give you Secretary
 2   Norm Mineta.
 3            KEYNOTE ADDRESS - HON. NORMAN Y. MINETA,
 4                      SECRETARY OF COMMERCE
 5             SECRETARY MINETA:  Greg, thank you very, very
 6   much, and I want to thank you for the terrific job that
 7   you are doing as our both head of Communications and
 8   Information as well as your dual other job as
 9   Administrator of the NTIA.  
10             First of all, I want to thank all of you for
11   taking time from your own busy schedules to be here with
12   us today.  This is a very, very important conference and
13   I'm really impressed with all of you being here.
14             Two years ago when we held a privacy summit, a
15   poll was released that said that more than 80 percent of
16   Americans were concerned about their threats to their
17   privacy when they were on line.  Unfortunately, according
18   to recent reports, the numbers have not changed and a Pew
19   study shows that over 86 percent of Internet users are
20   still concerned about businesses or people they don't know
21   getting access to their personal information.
22             Now, what has changed, however, is this, and
23   that is the way that industry is responding to these
24   concerns and the new technologies that are available to
25   protect online privacy.  That's why we are all here today,
0005
 1   to explore what those might be.
 2             In 1997, when a fraction of today's web
 3   population was on line, President Clinton and Vice
 4   President Gore issued their policy framework on e-
 5   commerce.  They understood that if e-commerce were to
 6   succeed it would be under private sector leadership.  It
 7   was the government's role to foster an environment to
 8   promote e-commerce and to work with the private sector to
 9   ensure that the Internet would grow worldwide. 
10             Since then the Internet without doubt has
11   flourished and technology has dramatically lowered costs. 
12   More and more people and businesses are online and e-
13   commerce is projected to climb into the trillions of
14   dollars over the next several years.  Now, all of us, both
15   in the private sector and on the government side, want
16   this to prosper.
17             The Internet has been a central part of the
18   longest peacetime economic expansion in our history.  But
19   we don't think this growth will continue unless both
20   consumers and businesses are confident about their
21   experiences on the net.  So we see privacy as a make or
22   break issue.
23             The administration is committed to continuing to
24   protect privacy and we have supported legislation to
25   protect children, sensitive financial information, medical
0006
 1   and genetic information as well.  And we have worked
 2   closely with the private sector to implement meaningful
 3   self-regulatory privacy regimes.
 4             Now, there has been progress.  The FTC found
 5   that 88 percent of the Internet sites now have some form
 6   of disclosure.  We now have several third party seal
 7   organizations that certify a web site is complying with
 8   its privacy policy, and a number of companies working with
 9   my predecessor Secretary Bill Daley prompt other companies
10   to adopt and disclose good privacy policies.
11             Now, these companies use their market leverage
12   by withholding advertising from web sites that failed this
13   test.  But it is a long road that we are all traveling,
14   and it seems to me much work remains to be done.  In the
15   FTC survey, the Federal Trade Commission survey, only 20
16   percent of Internet sites had policies that tracked fair
17   information practices.
18             Now, clearly we have a challenge, which I think
19   can best be met if we all work together -- government,
20   industry, and consumers.  One, we need to be sure that all
21   content providers satisfy the fair information principles. 
22   Among other things, they must provide Internet users with
23   information about who is collecting their personal data,
24   how it will be used, and how its  disclosure will be
25   limited.
0007
 1             Secondly, we need to empower consumers so that
 2   they can make informed choices about data collection. 
 3   Internet users should have a choice about whether to share
 4   information and with whom they want to share it.
 5             Three, consumers need to have reasonable access
 6   to information collected about them and an effective mans
 7   of recourse.
 8             Finally, we are here today to let consumers know
 9   what tools are available to protect their data.
10             Now, industry is doing a good job in developing
11   privacy-enhancing technologies, but the word hasn't gotten
12   out to the consumer.  Reports show that only one in 20
13   Internet users have used software that hides their
14   computer identity from web sites.  Only 10 percent of all
15   users have set their browsers to reject cookies.
16             Now, there are a number of privacy enhancing
17   technologies on the market or in the final stages of
18   development.  Many incorporate the Platform for Privacy
19   Preferences and this allows users to determine if a web
20   site meets their privacy preferences.
21             Let me indicate that all of us here at the
22   Department of Commerce are committed to making our web
23   site compliant with these standards.  Now, online privacy
24   is a critical issue for Internet users, one that we all
25   have to address if the extraordinary potential of this
0008
 1   technology is to be fully realized in the twenty first
 2   century.  So we have invited all of you here today to put
 3   a spotlight on privacy-enhancing technologies and what
 4   their capabilities are and to learn from you what might
 5   stand in the way of their future deployment and use.
 6             Tomorrow on Capitol Hill, Senator Hatch and the
 7   Internet Caucus will be hosting similar demonstrations to
 8   further highlight this issue.
 9             Let me close by saying thank you very, very much
10   to all of you for taking the time to be here and to
11   participate, and I look forward to your comments and, more
12   importantly, your recommendations on how we can enhance
13   privacy online and make sure that the Internet is a secure
14   place to visit.
15             Thank you very, very much.
16             (Applause.)
17             MR. ROHDE:  Mr. Secretary, thank you very, very
18   much.  We know that the Secretary's schedule is very, very
19   tight and we are very pleased that he was able to be here.
20             Before we begin the panels, we wanted to provide
21   a general overview of privacy on the Internet.  To do so
22   we have asked Dr. Lorrie Faith Cranor of AT&T Labs to do a
23   general presentation that will provide kind of a backdrop
24   for the discussion we will have throughout the day.  Dr.
25   Cranor is a senior technical staff member at AT&T and has
0009
 1   done a great deal of research that's focused on a variety
 2   of areas where technology and privacy policy intersect,
 3   including she's worked in areas of online privacy,
 4   electronic voting, and spam.
 5             She is currently the Chair of the Platform for
 6   Privacy Preferences Project Specification Working Group
 7   and the Co-Chair of the P3P Interest Group at the World
 8   Wide Web Consortium.
 9             So I would like to introduce for you Dr. Cranor.
10             (Applause.)
11       PRESENTATION:  OVERVIEW OF ONLINE PRIVACY-ENHANCING
12        TECHNOLOGIES - LORRIE FAITH CRANOR, AT&T RESEARCH
13             DR. CRANOR:  Good morning.  I'm going to try to
14   take you through kind of a whirlwind tour of the different
15   types of privacy technologies which are currently
16   available and show you some samples of what they look like
17   and give you a brief overview of how they work.
18             As has already been mentioned, we know that
19   online privacy is important because it appears in the
20   comics.  In this particular comic strip of "Cathy," which
21   had about a two-week series on online privacy, in this one
22   Cathy is finding out just how much can be found out about
23   her online.  In this conversation she learns that Irving
24   has found out the following.  He says:  "You love old
25   movies where men are heroes, love is forever, and women
0010
 1   get to wear little hats."  And she says:  "You know that
 2   about me?"
 3             He says:  "Yes.  You dream of snuggling up under
 4   a vintage quilt, sipping cocoa from Southwestern mugs,
 5   listening to the Three Tenors on CD."  She says:  "Irving,
 6   you, you created an online profile of me."
 7             He says:  "Yes, and also your stress-relief
 8   tablets are back ordered.
 9             So Cathy wasn't too pleased about this and in
10   the next few strips we learn how Irving was able to find
11   all this information out about her, and we find out that,
12   first of all, he snooped her e-mail and he looked at the
13   files on her computer because her employer had actually
14   hired Irving, and he was able to observe the chatter sent
15   by her browser, all the extra information that her browser
16   sends, and he was also able to set cookies through banner
17   ads and to use web bugs that allowed him to track her
18   activities across web sites. 
19             (Screen.) 
20             So let's take a minute to look at what browser
21   chatter is.  Browser chatter is basically all the things
22   that your browser is saying when you make little requests
23   on the web.  This will include your IP address, your
24   domain name.  It includes the referring page, and that's
25   the web page that you visited before the page you're
0011
 1   currently at.
 2             It also has information about your computer,
 3   including your operating system, what kind of browser it's
 4   using.  It also includes information about why requests
 5   you're actually making, so the URL, and if you're
 6   performing a search what are the search terms that you're
 7   using.  And it also includes cookies.
 8             Now, who gets to hear this chatter?  Well,
 9   obviously the web site you're visiting gets to hear this,
10   but there are other people who might hear it as well. 
11   This includes your local system administrators, your ISP. 
12   There may be other third parties involved, such as
13   advertising networks.  And any of this information that
14   may end up in log files at any of these places might be
15   subpoenaed later.
16             (Screen.)
17             So here is an example of the typical request,
18   and this is an actual request that I did.  I actually
19   shortened it a little bit so it would fit on the screen. 
20   But I went to buy.com and I wanted to buy some beer, and
21   so I did a search for beer and this is what my browser
22   transmitted.
23             You can see here that it includes my actual
24   request for beer, it includes the fact that I am
25   communicating in English, that I'm from the United States,
0012
 1   what web site I was at before I did the search, as well as
 2   a cookie which right now indicates that I have an empty
 3   shopping basket.  But there is all sorts of fields where
 4   they could put more information in there. 
 5             Since I performed this search on my employer's
 6   network, they know that I was shopping for beer at work.
 7             (Screen.) 
 8             Okay, what about cookies?  We've heard a lot
 9   about cookies in the past and cookies are actually a very
10   useful thing.  I like to think of them kind of like
11   staples.  If you're going in the physical world and you're
12   filling out a form and the form is multiple pages, often
13   you have multiple sheets of paper that are actually
14   stapled together.  And if it's a well-designed form, you
15   only have to write your name on the front page of the
16   form.  You don't have to write it on every page because
17   the form is stapled together and they know that all the
18   information on page 2, 3, and 4 applies to the same person
19   who wrote their name on page 1.
20             Well, cookies can perform that sort of function
21   in the online world, basically tieing together information
22   that is submitted on multiple pages of a form.
23             Another use for cookies is to identify you when
24   you return to a web site so you don't have to remember a
25   user name and password.  And it also can be useful for web
0013
 1   sites to get a better understanding of how people are
 2   using the web site, so when you go to a newspaper site,
 3   does everybody read the sports page first; they'd like to
 4   know those sorts of things.
 5             So there are some very useful uses of cookies. 
 6   But as we've heard, there are also some reasons where
 7   cookies can be harmful.  In particular, people are
 8   reacting to cookies being used to profile users without
 9   their knowledge.  So they can be used to monitor users
10   across multiple web sites in the background, so that users
11   don't even know that they're being tracked. 
12             (Screen.) 
13             An example of how this can work.  So imagine
14   that you're going to a search engine and you're doing a
15   search for, say, some medical information and the search
16   engine has an ad, and that ad is going to set a cookie on
17   your computer.  When it sets a cookie on your computer,
18   basically it's sending this little bit of data to your
19   computer, and whenever you go back to the web site that
20   set the ad your computer is going to send that little bit
21   of data back.
22             In this case, it's not the search engine that
23   set the cookie, but it's the ad company.  So after I've
24   done my search, now I'm going to go to a book store and
25   buy a book.  It turns out that this book store uses the
0014
 1   same ad company as the search engine, and so when it goes
 2   to display the ad it's going to send the cookie back to
 3   the ad company.  So now the ad company knows that I went
 4   to the search engine and then I went to the book store. 
 5   It has tracked me across multiple sites.
 6             If the ad company has the cooperation of the
 7   search engine and the book store, it may actually be able
 8   to get some more information, because while the cookie
 9   itself doesn't know who I am, when I bought that book I
10   had to provide my name and my address and so the book
11   store knows who I am, and if the book store cooperates
12   with the ad company now they have an identified profile of
13   me.
14             Another thing that you may have heard about in
15   the media is web bugs.  Basically what a web bug is, it's
16   very similar to what I just showed you, where you have an
17   ad that's sending a cookie, except that a web bug is
18   invisible.  So there's some invisible little dot on your
19   screen that has a cookie associated with it and it may be
20   tracking you, but not only do you not know that's
21   happening, but you can't even see that the dot is there.
22             (Screen.) 
23             Other privacy problems have come up due to the
24   refer.  As I mentioned, the refer is the address at the
25   last web site you visited before the current one.  Often
0015
 1   when you go to web sites that have forms, sometimes after
 2   you fill out the form, if you look carefully at the URL
 3   bar, you'll notice that it changed and that some of the
 4   information you typed into the form is now actually in
 5   that URL.  So there's an example at the bottom of the
 6   screen where you have an actual person's name and address
 7   that actually becomes part of that URL.
 8             Now when I go to the next web site, that
 9   information is going to be transmitted to that web site. 
10   So this is a big problem.  Now, there are a lot of
11   companies that once they became aware of the problem they
12   changed the way their web forms worked so that doesn't
13   happen.  But there still are a number where this is still
14   a problem. 
15             (Screen.) 
16             Okay, so what can you do about this?  Well, this
17   is a slide actually that I borrowed from one of my
18   colleagues that he put together.  He says, well, you know,
19   you can go to cyber cafes, only use the Internet while
20   you're drinking coffee at different cyber cafes, and of
21   course never go back to the same one.  And you can use
22   free e-mail services instead of an ISP and keep changing
23   them, or set up a prepaid cash account with your ISP and
24   be sure to give all phony information.  And you can forge
25   e-mail, and of course never go shopping online where you
0016
 1   have to actually give out personally identifiable
 2   information.  And if you do all that, then you should be
 3   pretty safe.
 4             Fortunately, there are some other solutions. 
 5   There are a number of software tools, many of which you
 6   are going to see demonstrated today in the foyer, that can
 7   help people address some of these concerns.
 8             The first set of tools I'm going to talk about
 9   are anonymity and pseudonymity tools, which basically help
10   people surf the web without any of their actions being
11   linked directly to them.  Another set of tools is
12   encryption tools and I'm going to touch on them briefly. 
13   That's not really the focus of the session today.  We
14   could do a whole another session just on encryption tools. 
15   Then another set of tools are filters.  Basically, these
16   are tools that are going to be making sure that your
17   computer is not sending out certain kinds of information
18   and so they may be blocking cookies, they may be making
19   sure that your child is not sending out their name or
20   their phone number over the Internet, things like that. 
21             (Screen.) 
22             There are also some tools that are helpful for
23   information and transparency, so they basically inform
24   users as to what's going on, what information is collected
25   about them, and what is going to happen to it.  Then there
0017
 1   are a whole bunch of other tools that sort of didn't
 2   really fall neatly into these other categories, and I'm
 3   going to talk about a few of those. 
 4             (Screen.) 
 5             Just to give you sort of a holistic picture of
 6   how this all works, you can imagine that you have your
 7   user and you have various services and web sites and there
 8   is the Internet that's sort of this big cloud in the
 9   middle, and the user lives somewhere where there's a
10   regulatory and self-regulatory framework and the services
11   that they visit may be following the same laws and the
12   same guidelines or they may be following different ones.
13             Then the user has the ability to send all their
14   messages on the Internet through a number of different
15   tools.  So these might include an anonymizing agent or a
16   cookie cutter or a P3P tool.  So the user can basically
17   get extra privacy features by using these tools. 
18             (Screen.) 
19             Then the other thing is that it's very important
20   that any data that the user does decide to send to a web
21   site is going to be visible only to that web site.  So the
22   user needs to have a secure channel through encryption
23   tools to make sure that this data is just not picked up by
24   anybody else who happens to be listening. 
25             (Screen.) 
0018
 1   
 2             So let's start with anonymizing proxies.  This
 3   is a tool that basically a user can set up their web
 4   browser so that all the requests they make to the Internet
 5   are going to go through this proxy server.  Basically, the
 6   proxy server takes the request, strips off identifying
 7   information, and forwards it to wherever the user wants it
 8   to go.
 9             Then when the web site responds, it responds by
10   sending the request back to the proxy.  The proxy knows
11   what user requested it and so they can send it back to the
12   user.  So as a result, the end servers have no idea who
13   the original user is.  However, the proxy can see
14   everything and so we have to trust that the proxy is going
15   to have a good policy and not be using that data in ways
16   that the user doesn't want.  There are a number of
17   different proxy-based services.  Some are free, some are
18   subscription, some are supported by advertising. 
19             (Screen.) 
20             One of the best well-known ones is called the
21   anonymizer, anonymizer.com.  For example, if you are
22   visiting Yahoo using the anonymizer it would add this
23   little blue bar at the top of your screen to indicate that
24   the page was loaded by anonymizer.  It includes a control
25   panel that lets people control some of the settings, for
0019
 1   example whether or not you want the anonymizer to filter
 2   out the cookies or let them through.  There are a few
 3   other settings that the user can control. 
 4             (Screen.) 
 5             There are also a number of tools that are
 6   related to anonymity tools, but they have sort of an extra
 7   thing.  I call them pseudonymity tools.  They have the
 8   ability to automatically generate user names, passwords,
 9   e-mail addresses, or other information, basically
10   pseudonyms for the user, and keep track of them.  A lot of
11   users say that when they go to web sites and they're asked
12   to provide information that they lie and they make up
13   names.  But actually, you have to keep track of what it is
14   you made up if you want to maintain a relationship with
15   the site.
16             So these tools will actually keep track of that
17   automatically for the user.  So one example is iPrivacy,
18   and I believe they're going to be doing a demonstration
19   here.  Basically, when I go to a web site and I want to
20   order something and have it shipped to me, I can type in
21   my information and iPrivacy will translate that
22   information into this essentially encrypted private
23   identity and then it will fill out the form for me with
24   this translated information, and then the web site will
25   get the encrypted information.
0020
 1             They will have enough information to give to the
 2   credit card company and to give to the shipping company so
 3   that the transaction can be processed, but they web site
 4   itself doesn't know what my credit card number is or what
 5   my home address is, and in the end the shipping subsystem
 6   can basically print an address label which will allow the
 7   product to be delivered to me without the company actually
 8   knowing who I am or where I live. 
 9             (Screen.) 
10             This is the screen.  They have to download some
11   software onto the computer to make this work. 
12             (Screen.) 
13             There's another company called Incogno which has
14   a similar tool, except it doesn't require that users
15   download it onto their computer.  They work with
16   merchants, and if the user visits a web site that is
17   equipped with Incogno they can use a similar tool where
18   Incogno may have access to some of this personal
19   information but it is not actually released to the vendors
20   themselves. 
21             (Pause.)
22             Now, this is where technology breaks down. 
23   Let's see if I can revise this.  Otherwise I'll wing it. 
24             (Screen.) 
25             Now we're in trouble.  
0021
 1             Okay, while my computer is rebooting, some of
 2   the other tools that I'd like to show you pictures of, but
 3   you'll just have to imagine, besides the anonymity and
 4   pseudonymity tools, the next tools I was going to talk
 5   about are encryption tools.  There are encryption tools
 6   which are useful to encrypt data in a variety of ways. 
 7   There are some which are focused on encrypting data as
 8   it's being transmitted from your computer to the web site. 
 9   There are also tools which are focused on encrypting e-
10   mail messages.  And there are some which are focused on
11   encrypting the data when it's on your computer or when
12   it's on the database out there. 
13             I'm not really going to talk about those in
14   detail, other than just to say that they're really
15   important.  If you have all of these other privacy tools
16   and your data is just floating out there unencrypted,
17   that's not really going to offer a whole lot in the way of
18   privacy protection.  So it's important not to leave out
19   those things.
20             Now, fortunately, increasingly some of these
21   tools are actually being built into web browsers, being
22   built into database systems, so that's going to be very
23   helpful.
24             The next set of tools that I wanted to talk
25   about are the filtering tools and the first set of
0022
 1   filtering tools are the cookie cutters.  So these are
 2   tools which can be configured to block all cookies or to
 3   selectively block cookies, and often users will want to
 4   allow cookies from some web sites where they see
 5   themselves as actually getting some value from having the
 6   cookie, and at other web sites they'd rather not have the
 7   cookie. 
 8             Some of these tools also allow the users to go
 9   ahead and accept the cookies and then to review them later
10   and decide which ones they want to keep and which ones
11   they want to discard.
12             Besides being able to filter cookies, a lot of
13   these tools also have the ability to filter out the refer
14   header and some of the other browser chatter that we
15   talked about.
16             Hang on a second.  We'll be back with the
17   presentation. 
18             (Pause.) 
19             (Screen.) 
20             Okay.  The other kind of filter besides the
21   cookie cutter are the child protection software.  Now,
22   most of this software was developed primarily to allow
23   parents to filter out material that they felt was
24   inappropriate for their children.  But a lot of this
25   software also has a feature where parents can block the
0023
 1   child from sending privacy-sensitive information on the
 2   Internet.  So for example, they can set it up so that the
 3   child's name or phone number won't be sent on the
 4   Internet.  They can also limit who a child can e-mail or
 5   chat with. 
 6             (Screen.) 
 7             Another type of tool that we talk about are
 8   identity management tools.  Sometimes these are referred
 9   to as infomediaries.  Some of these companies refer to
10   themselves as infomediaries and some of them tell me that
11   they absolutely aren't infomediaries.  So I'll just call
12   them identity management tools.
13             But basically, the idea is that these are tools
14   that help people manage their online identities. 
15   Basically, they generally offer some sort of an electronic
16   wallet or electronic archive where I can type in my
17   information in some sort of a secure storage and have the
18   ability to have my computer automatically send this
19   information when I authorize it.  So they'll automatically
20   fill out forms or automatically send demographic
21   information to web sites, but only when I have authorized
22   it.
23             So some of them are essentially an opt-in to
24   targeted advertising.  Some of them will pay consumers for
25   data.  Some of them actually go and check for privacy
0024
 1   policies at web sites and give consumers some indication
 2   as to what kind of privacy policy the site has before they
 3   will release the data.  There are a number of different
 4   examples.  They all have slightly different models of how
 5   this actually works.
 6             One example is Persona.  In Persona, the
 7   consumer will in advance fill out this profile with
 8   personal information and they will indicate for each field
 9   when they want to allow it to be shared.  And then Persona
10   Valet has this tool bar that will allow users to have some
11   control over cookies and control over when to provide
12   data, and they are also planning on building in P3P so
13   that they can give users alerts about web sites' privacy
14   practices before data is released. 
15             (Screen.) 
16             There's another company called Privacy Bank that
17   has an interesting system.  Here if I go to a web site
18   that's equipped with Privacy Bank -- in this case this is
19   the Starbucks Coffee site -- if I want to order some
20   coffee beans, I will click on my Privacy Bank bookmark
21   which I set up when I subscribed to Privacy Bank, and it
22   will pop up this window which provides a snapshot of what
23   the Starbucks privacy policy is.  There are little symbols
24   that you probably can't see, but they basically give me -
25   - there are five symbols and depending on which ones
0025
 1   appear I have some idea of what their privacy policy is.
 2   
 3             Then if I'd like to fill out the form, then I
 4   can click on the "My information" and drag it onto the
 5   form, and at this point it gives me an alert, because when
 6   I registered for Privacy Bank I indicated what my personal
 7   preferences are.  There's a conflict here, and so it says
 8   "This site does not meet your privacy preferences.  Would
 9   you still like to fill out the form?" 
10             I can say yes or no.  I can also click on the
11   button for policy details and get a much more detailed
12   indication of what kind of data they collect and what
13   exactly they do with it, and then I can make an informed
14   decision. 
15             (Screen.) 
16             The next thing I'd like to talk about is the
17   Platform for Privacy Preferences, or P3P, and I'm just
18   going to give you a brief overview of that and that will
19   be discussed more on the next panel.  But basically, P3P
20   is designed to give web sites an easy way to take their
21   privacy policies and convert them into a standard machine-
22   readable format.  Then once this is done, we can build web
23   browsers and plug-ins and other tools that can
24   automatically go fetch these machine-readable policies and
25   read them for the user, and then they can compare the
0026
 1   policies with the user's preferences and alert the user
 2   when there are conflicts. 
 3   
 4             (Screen.) 
 5             So basically a web site that wants to use P3P,
 6   it's fairly straightforward.  First they have to have a
 7   privacy policy.  Then they need to translate it into P3P
 8   format, and there are a number of tools that are available
 9   to assist with the translation process.  Then they take
10   that translated policy and they put it on their web site,
11   and then they create another policy -- another file which
12   indicates what parts of the web site this policy applies
13   to.
14             So a web site can say:  We have one P3P policy
15   for the entire site, or they might have different policies
16   for different parts of the site and they can indicate what
17   policy applies where. 
18             (Screen.) 
19             The P3P vocabulary basically has a number of
20   different fields which capture what we felt were the
21   important parts of a privacy policy that people would be
22   interested in knowing.  So basically they are who is
23   collecting data, what data is collected, what purpose will
24   it be used, is there an ability to opt in or opt out, who
25   are the data recipients, what kind of access is provided,
0027
 1   what kind of data retention policy is there, how will
 2   disputes about the policy be resolved, and that includes
 3   third party seals, laws, anything along those lines, and
 4   then finally where is the full human-readable policy if I
 5   want to go get more information. 
 6             (Screen.) 
 7             One of the advantages of P3P is that not only
 8   can I easily find out about a web site's privacy policy,
 9   but I can also find out if the web page has any other
10   objects in it that might have different privacy policies. 
11   So for example, here you see this page which is on the
12   AT&T web site and as a user I look at it and it looks like
13   it's just one whole page.  But in fact it actually has an
14   ad in it and this ad is served from another company, and
15   the privacy policy associated with that ad is the policy
16   of that other company.
17             So using P3P, my user agent should be able to
18   automatically discover this and check both privacy
19   policies to make sure they match my preferences. 
20             (Screen.) 
21             I'm going to show you a prototype P3P user agent
22   which was designed at AT&T in conjunction with Microsoft. 
23   This is a plug-in which is designed to work with Internet
24   Explorer, and it allows users to configure their
25   preferences.  This is the preference configuration screen,
0028
 1   which I don't expect you to actually read here, but you
 2   can see it's a fairly small number of questions.  The
 3   users can use that to specify what their personal
 4   preferences are.
 5             After they have set their preferences, whenever
 6   they go to a web site they can click on the privacy button
 7   and it will check the results.  So here are examples at
 8   two different web sites.  The top web site, there is a
 9   warning that comes up because there was a mismatch with
10   the user's preferences and it says:  "This site does not
11   allow you to find out what data they have about you." 
12   Basically, the site provides no access and I have said
13   that I need access.
14             Then the bottom screen, we're visiting a site
15   where there are no warnings.  We can see here that this
16   site has a seal from TrustE.  That's one of the things
17   that comes up as part of a policy. 
18             Now, if a user changes their privacy preferences
19   and goes back to that same site, now all of a sudden they
20   have a warning that this site may collect data that does
21   identify you for profiling.  So basically now the user has
22   said:  Hey, I actually don't really want profiling, and so
23   that warning will come up. 
24             (Screen.) 
25             Another P3P user agent is by IDcide.  It's
0029
 1   called the Privacy Companion and it's a plug-in that works
 2   for both Internet Explorer and for Netscape, and it
 3   provides cookie management capabilities.  It actually
 4   started out as a cookie management tool and by adding P3P
 5   to it they've been able to enable people to have much more
 6   fine-grain control over when to allow cookies.
 7             So it adds these little symbols to the top of
 8   the browser bar and the symbols indicate red for when
 9   there's a P3P policy that's not acceptable, green when it
10   is acceptable, and grey when there's no policy at all. 
11   Then when I visit a site that's trying to set a cookie, if
12   the privacy policy doesn't match I can get information as
13   to exactly why that pops up here.
14             Another tool is called YOUPowered Orby Privacy
15   Plus, and this is actually a toolbar that sits on the
16   user's desktop rather than in the browser window, and it
17   has similar features.  It also has this trust meter.  You
18   can see in the upper right-hand corner of the screen, it
19   gives you a little visual graphic indication of how close
20   the policy is to matching your preferences, and then you
21   can click on it to get specific positive and negative
22   flags to see specifically where the web site may have
23   problems or where it's doing good things.  You can also
24   get it to prompt you when web sites want to set a cookie
25   and they don't necessarily match your preferences. 
0030
 1             (Screen.) 
 2             IBM has a tool that they're demonstrating here
 3   that lets web sites create their P3P policies.  This is
 4   something that's currently available from the IBM web site
 5   and it's a drag-and-drop interface that basically somebody
 6   can pick up these little elements on the screen which
 7   represent different kinds of data and drag them onto the
 8   other page to indicate we collect this kind of data, and
 9   then they can fill out various information about each type
10   of data and what is done with it.
11             IBM also has a number of templates, so rather
12   than starting from scratch you can find -- 
13             (Screen.) 
14             -- a template that follows similar practices as
15   your web site and then just edit that to customize it for
16   your web site. 
17             (Screen.) 
18             Another tool which allows web sites to create
19   P3P policies is called PrivacyBot.  They're actually a
20   web-based interface.  I go to their web site as a
21   webmaster and I can fill out their lengthy questionnaire
22   about my web site and I can pay I think it's $30 with my
23   credit card and create both a P3P policy as well as a
24   human-readable privacy policy for my web site. 
25             (Screen.) 
0031
 1             Finally, YOUpowered, which I showed you a minute
 2   ago, also has a tool for webmasters to create their
 3   privacy policies as well. 
 4   
 5             (Screen.) 
 6             Now I want to mention a few other tools.  As I
 7   said, these are sort of the miscellaneous ones.  There are
 8   some privacy-friendly search engines.  One example is
 9   TopClick.  These are basically search engines which are
10   committed to not using cookies, not tracking users, and
11   basically trying to be as privacy-friendly as possible.
12             Another type of tool are computer cleaners. 
13   When you're surfing the web there's all sorts of little
14   files that are created on your computer to kind of keep
15   track of things while you're surfing and many of these
16   files are no longer really needed once you're done.  So
17   there's a tool called WindowWasher which goes through a
18   removes all of these files, therefore removing the traces
19   of what web sites you visited when you were on line. 
20   People say that it actually makes your computer run
21   faster, too. 
22             (Screen.) 
23             Another type of tool is tools to facilitate
24   access.  We're going to hear today from a company called
25   Privacy Right, which is focused specifically on access. 
0032
 1   So for example, they have, for example, this tool.  I can
 2   go to a web site that is equipped with their tool and
 3   indicate what kinds of data uses are acceptable, basically
 4   opting in and out of various things, and then I can
 5   actually go and view what data they have collected on me
 6   and who it has been disclosed to and when, and I can view
 7   that on line.  I can also request that they send me a
 8   complete audit trail of where my data has gone.
 9             So that's pretty much the end of our whirlwind
10   tour.  I think I even managed to stay mostly within time. 
11   The final thing that I want to leave you with is that
12   we've seen a variety of different tools, but there's no
13   one tool which is going to solve all the problems.  Really
14   what we need is tools that can work together.  So I think
15   each tool has its strengths and it has something that it
16   can contribute.
17             So P3P tools are useful to help users understand
18   privacy policies, but they don't deal with, for example,
19   enforcing privacy policies.  Seal programs can help with
20   enforcement as well as regulations can also help with
21   enforcement.  Anonymity tools and filtering tools can
22   reduce the amount of information that I actually reveal,
23   which is great when I don't need to reveal information,
24   but if I want to actually go purchase something there may
25   be information that I do need to reveal, and so then some
0033
 1   of the other tools are going to be useful.
 2             Encryption tools are of course useful for
 3   securing data both in transit and in storage.  Finally,
 4   the laws and codes of practices are going to do what tools
 5   can't, which is provide a baseline level of acceptable
 6   policies.
 7             Thank you.  And if you would like a copy of any
 8   of these slides, they are all available on my web site.
 9             (Applause.)
10             MS. LEVY:  Thank you so much, Dr. Cranor, for
11   the extremely informative overview.
12             We're now ready for our first panel of the
13   morning, P3P Implementation.  I ask all the panelists to
14   come up.
15              PANEL DISCUSSION:  P3P IMPLEMENTATION
16             MS. LEVY:  Our moderator for this panel is
17   Elliot Maxwell, the Special Adviser to the Secretary of
18   Commerce for the Digital Economy.  Among other things,
19   Elliot advises the Secretary on and helps coordinate web
20   site of Commerce activities regarding electronic commerce
21   and the Internet.  I'm going to ask that Mr. Maxwell
22   introduce his own panelists.
23             MR. MAXWELL:  We are under this time pressure
24   and we are just going to hustle.  So Lorrie just did a
25   wonderful job.  It's like having a whole sort of panel to
0034
 1   herself, but giving us an enormous amount of information
 2   about privacy in general.  We're going to focus right now
 3   in this panel on P3P.  
 4             I will give a sort of blast fax introduction of
 5   the panelists and then they will have three and a half
 6   minutes.  You will see a clock counting down at the end,
 7   so the last 30 seconds you count down like it's a NASA
 8   countdown, so they'll know that they don't have any more
 9   time.
10             We're going to try to have questions at the end,
11   so we'll try to go very quickly.  If we are able to do
12   that, there are microphones in the aisles, so that if you
13   will use those we can have a recording of it.
14             I will give the introductions in order of, in
15   alphabetical order, though that will not match the order
16   of appearance, and I'll give that quickly at the end.  But
17   let me just briefly introduce the panel:  Marc Berejka is
18   a Senior Federal Affairs Manager and Senior Corporate
19   Attorney at Microsoft's corporate office in Washington,
20   responsible for developing and advocating the company's
21   positions in telecommunications issues and in particular
22   in this regard in the area of privacy.
23             He chairs the Information Technology Industry
24   Council's working group on telecommunications policy and
25   the Information Technology Association of America's
0035
 1   working group on information policy.
 2             Karen Coyle is a digital library specialist at
 3   the University of California.  She speaks and writes often
 4   on issues relating to cyberspace and is active with
 5   Computer Professionals for Social Responsibility, a public
 6   interest group based in Palo Alto.  She's written
 7   critically about P3P, which she refers to as "pretty poor
 8   privacy," so she carries a particular perspective on this.
 9             Dierdre Mulligan is Staff Counsel at the Center
10   for Democracy and Technology and works with groups around
11   the world on fair information practices and strengthening
12   individual's control over their personal information.
13             Dan Jaye is CTO and Co-Founder of Engage, part
14   of the CMGI Group, where he is involved in delivering
15   interactive database marketing and information service
16   products.  He's also involved with a group itself in
17   facilitating communications about technology and privacy.
18             Ed Mierzwinski has been a consumer advocate with
19   the Public Interest Research Group, U.S. PIRG, since
20   January of 1989.  He's a frequent participant in public
21   policy forums, has testified often at both the state and
22   federal levels, and is a member of the steering committee
23   of the Trans-Atlantic Consumer Dialogue, which just
24   recently met in Brussels.
25             Ron Perry of IDcide is the company's CEO. 
0036
 1   IDcide is a company that provides a bridge between privacy
 2   and profiling services to Internet users and tries to
 3   address the privacy needs of both e-businesses and
 4   consumers.  They just released "The Privacy Companion,"
 5   which is a product that reports cookie-based tracking,
 6   automatically blocks third party cookies, and eliminates
 7   data spillage problems.
 8             Mel Peterson of Proctor and Gamble has served in
 9   a number of different positions with the company.  In 1998
10   he joined the Interactive Team at P and G -- it's now
11   called I-Ventures -- and he helped implement P and G
12   systems to manage online privacy and to develop P and G's
13   global privacy guidelines.  Effective this year, just
14   about a month ago, Mel became P and G's global privacy
15   manager, responsibility for building and enhancing P and
16   G's privacy management capabilities globally.
17             Martin Pressler of IBM is an advisory programmer
18   at IBM's facility in Research Triangle.  He's been deeply
19   involved in P3P and is co-author of the P3P specification.
20             Danny Weitzner, on his right, is Director of the
21   World Wide Web Consortium's Technology and Society
22   Activities, responsible for the development of technology
23   standards that enable the web to address social,
24   political, and public policy concerns, and has been again
25   heavily involved in P3P.  Before joining W3C he was at the
0037
 1   Center for Democracy and Technology and was also at EFF.
 2             So I think I got my three and a half minutes. 
 3   You won't hear much from me any more.  So Danny, why don't
 4   you take it away.
 5   
 6             MR. WEITZNER:  Thanks very much, Elliot.
 7             Because we have a big issue and a big panel, I'm
 8   going to speak in sort of a cryptic, skeletal way.  I
 9   really just want to highlight, following Lorrie's
10   presentation, which I think gave us all a good background
11   on what P3P is and how it works, I want to just make three
12   very quick points about why we need P3P on the web, and
13   three points about why I think it's actually going to
14   happen.
15             First of all, I think there's a tremendous need
16   for what Lorrie referred to as machine-readable privacy
17   policies.  That is, anyone who looks out on the web today
18   and looks at privacy policies knows that they're
19   relatively long, they're written, even for lawyers, in
20   sometimes somewhat confusing language, and I think that
21   some tend to feel that this is maybe done to confuse or
22   deceive users. 
23             I actually think to a large extent it reflects
24   the fact that privacy practices, the practices of handling
25   personal information on the web, are becoming increasingly
0038
 1   complex.  Therefore, I think we desperately need to put
 2   privacy policies in machine-readable format, in the same
 3   way as everything else on the web is in machine-readable
 4   format, so that the tools that we use to access the web,
 5   the browsers and all the other user agents out there, can
 6   help us to decipher these policies and make intelligent
 7   decisions about them.
 8             Secondly, I think that we badly need P3P because
 9   we need to harness the entrepreneurial energy that has
10   made the web itself work to help users make intelligent
11   decisions.  To me one of the most extraordinary things
12   that's happened in the development of P3P is that, now
13   that the standard is more or less complete, we're seeing
14   the whole range of companies come out with a variety of
15   products that give users a variety of different options
16   about how to make choices about their personal
17   information.
18             I think it's safe to say that the designers of
19   P3P didn't even think about a lot of these products or
20   about these possibilities.  But we need to provide that
21   kind of flexibility to users which in practice in the web
22   arena is provided when we let software developers go out
23   and do innovative things.
24             Third, I think that, for the reasons I've said,
25   P3P really, in the words of Alexander Dix, who's a data
0039
 1   protection commissioner from Germany, is necessarily but
 2   not sufficient to guarantee online privacy.  As Lorrie
 3   indicated, P3P is part of a broad array of tools,
 4   services, self-regulatory practice, regulation, that all
 5   have to come together to make privacy work.  But P3P is
 6   really necessary for that. 
 7             I think there are three reasons following those
 8   why web sites will adopt P3P, why software developers will
 9   build P3P products, and why we'll really see this kind of
10   enhanced user control over privacy on the web.  Number
11   one, web sites, particularly those web sites that are
12   trying to have a commercial relationship with users and
13   sell them something, very badly want to ensure that users
14   have a seamless browsing experience.  They don't want
15   users to be distracted by going off and having to find the
16   privacy policies and take however many minutes or hours it
17   takes to understand it.  Seamless browsing is critical.
18             Second, I believe that as the major browser
19   vendors begin to integrate P3P into their products, and
20   two of the major browser vendors, Microsoft and Netscape,
21   have announced plans to integrate P3P, users will come to
22   expect to see a P3P policy.  I doubt very much that many
23   more than about a tenth of one percent of users will ever
24   say, I want my P3P, but they will see, as you saw on
25   Lorrie's slides, an icon that's an icon that's in grey
0040
 1   instead of in green and they'll wonder what's happening.
 2             Third, I think that P3P will become a part of
 3   the web and needs to become a part of the web, because web
 4   services are fundamentally becoming more complicated and
 5   they are becoming more integrated into our lives in a
 6   variety of ways, and the fact of the matter is that as we
 7   live our lives, whether it's online or off, we share,
 8   trade, quite a lot of personal information, and users need
 9   a way to control that. 
10             Certainly anonymity has its place.  It's
11   critical for protecting basic human rights and civil
12   liberties.  But in the world that the web is increasingly
13   becoming part of, there is lots of personal information
14   moving around and users need control of it.
15             MR. MAXWELL:  Danny, we're going to move on to
16   the next person because autocracy is the rule here.
17             Marc Berejka is next from Microsoft, talking
18   about this from the viewpoint of the browser manufacturer.
19             MR. BEREJKA:  Thanks, and I will try to keep
20   this on schedule --
21             VOICE:  Microphone.
22             MR. BEREJKA:  Can you hear me now? 
23             If we get network connectivity, we'll have live
24   connections to show the tools that I'll show you
25   graphically here.
0041
 1             MR. MAXWELL:  Can we get that mike, please, Marc
 2   Berejka's mike, please. 
 3             (Screen.) 
 4             MR. BEREJKA:  So and Lorrie and Danny have said,
 5   there's a basic chicken and egg issue here in delivering
 6   P3P.  Microsoft looks forward to helping to deliver in the
 7   near term both a chicken and an egg.  We've been committed
 8   to P3P for some time.  Our first manifestation of that
 9   commitment or tangible manifestation was in the
10   development of a privacy statement generator.  Again, we
11   have that out in a mockup in the lobby, but if we have
12   network connectivity it'll be live. 
13             (Screen.) 
14             If we could go to the next slide, you'll see
15   that what's been sort of the longstanding beta of this
16   privacy statement generator, a beta that's been running
17   for 18 months now, has gotten about 20,000 companies to
18   use it and to walk through, to walk through a basic
19   questionnaire as to what the web site's basic information
20   practices are. 
21             What we do with the web site generator is try to
22   guide users to follow the fair information practices. 
23   This slide indicates that we're actually asking web site
24   operators to consider how they provide access.  At the end
25   of the process, you press the "Done" button and it
0042
 1   generates a sample privacy policy with some indicators as
 2   to where the web site operator might want to seek some
 3   further guidance or where we believe information might be
 4   lacking.
 5             Again, this longstanding beta, if you will, has
 6   gotten 20,000 companies plus to use it.  The game plan is
 7   to update this privacy statement generator so that it's
 8   compatible with the current version of P3P and our privacy
 9   statement generator that is compatible with the current
10   version of P3P would be released in the December to
11   January time frame.
12             This version that's up now is compatible with an
13   April or earlier, April '99 or earlier version of P3P, so
14   it's not quite up to snuff, but is nonetheless useful for
15   basic purposes.
16             In terms of the client experience, if we could
17   go to the next slide -- 
18             (Screen.) 
19             -- we also want to make that as simple as
20   possible.  Right now we have just, as Lorrie showed,
21   developed a basic tool, and you can see this better out on
22   the screens in the lobby, but what we're struggling with
23   is how to simplify the P3P experience for the end user. 
24   In this mockup we ask the individual whether they only
25   want to go to a site that collects data necessary for the
0043
 1   processing of a specific request, whether they only want
 2   to go to a site that does not reveal your identity, and
 3   whether they only want to go to a site that does not
 4   identify you for profiling.
 5             So we're trying to make these choices simple for
 6   the consumer so that the process of enabling the browser
 7   to read XML statements is not difficult. 
 8             If we could go to the next slide -- 
 9             (Screen.) 
10             The next process, as Danny and Lorrie pointed
11   out, is for the P3P-enabled browser to talk to the XML-
12   enabled privacy statement.  If we could go to the next
13   slide -- 
14             (Screen.) 
15             -- this is a sample -- it's similar to what
16   Lorrie showed -- of what gets spit back.  We ran this test
17   live against Microsoft.com and what came back was that
18   Microsoft.com, the fact that Microsoft.com is a bearer of
19   the TrustE seal, that there's one-click access to the
20   Microsoft.com privacy policy.  But you know what? 
21   Microsoft.com does collect, does use cookies.  It's
22   disclosed in their privacy policy that they do use
23   persistent cookies to do some level of tracking.
24             What we're really looking for over the course of
25   the next nine months or so, because this process is
0044
 1   complicated, is input from interested people who want to
 2   help us in this boiling-up process to make the consumer
 3   experience in implementing P3P easy.  The game plan is to
 4   release the privacy statement manager -- excuse me, the
 5   privacy manager for the consumer, integrated into the next
 6   version of the operating system, which is codenamed
 7   Whistler and which is due out some time next year.
 8             MR. MAXWELL:  Thanks very much, Marc.  This is a
 9   world record for 11 slides in under 4 minutes.  So we
10   appreciate it. 
11             Next, Ron Perry from the standpoint of those
12   people developing applications that try to serve
13   individuals and increase their control over privacy.
14             (Screen.) 
15             MR. PERRY:  Thank you, Elliot.  In compliance
16   with the government regulations you set out here regarding
17   timing, I'll rush through my three slides. 
18             I wanted to give you our perspective on how we
19   think P3P can help applications treat privacy better, give
20   you a little overview of what we've done and what we think
21   are the possibilities that would be opened up by P3P.  So
22   first of all, what we've done with P3P is to utilize it in
23   order to simplify decisionmaking by the user.  Instead of
24   having the user read through a complex privacy policy and
25   then decide whether cookies should be enabled for this
0045
 1   site or not, understanding how the site uses information
 2   collected, we are trying to automate this decision.  So
 3   the Privacy Companion with P3P reads the privacy policy
 4   for a site, analyzes it, tries to find out whether it
 5   matches the user's preferences or not, and decides whether
 6   to allow the cookie from the site or not.
 7             We also show the user, as Lorrie showed in her
 8   slides earlier, we give the user a visual indication of
 9   whether the policy is acceptable or not according to that
10   user's preferences. 
11             (Screen.) 
12             We see the major advantage of P3P in its opening
13   up of a wide range of possibilities for application
14   developers.  The first type of applications that we can
15   see using P3P are privacy-enhancing technologies, such as
16   ours.  But we also see many possibilities for other tools
17   like search engines that can take the privacy policy of a
18   site into consideration.
19             Various privacy policies -- the fact that a
20   computer application can do something with a privacy
21   policy really opens up the issue of privacy into a
22   competitive advantage, so that the major challenge we see
23   in the development of P3P is being able to actually
24   enhance its vocabulary to emphasize the good practices
25   used by various companies and really give them a
0046
 1   competitive advantage, turning privacy into a competitive
 2   advantage, something that is not possible today without
 3   such technologies.
 4             Thank you.
 5             MR. MAXWELL:  Thank you, Ron.
 6             Next will be Mel Peterson.  Proctor and Gamble
 7   is sort of the mother church of brand management and so
 8   someone who takes the job of worldwide privacy policies
 9   for a firm like that has a big challenge.  Mel.
10             MR. PETERSON:  We were sort of a guinea pig for
11   P3P a few months ago.  We participated in the June interop
12   demos, created a P3P policy statement based on our privacy
13   policy that we have displayed on the pg.com web site. 
14   Neither I nor the developer that worked on this had even
15   read the P3P spec ahead of time, so we were a pretty good
16   test of what does it take for a company to actually do
17   this.  And we didn't cheat and use one of the nice
18   generators that are out there, either.  We wrote the code
19   by hand.
20             The net was it was very easy to do and very
21   inexpensive.  Over the course of a couple of weeks, we
22   spent maybe three elapsed or three effort days to create
23   the P3P statement, learn what it took to do that, debug
24   it, put it up on the site.  So it's a very straightforward
25   thing for content providers and web sites to do. 
0047
 1             (Screen.) 
 2             I was asked to speak to, so what do you think
 3   other content providers and web sites are going to react
 4   to P3P?  I would make four points.  The first is simply
 5   that my suspicion is most web sites and content providers
 6   haven't spent a lot of time learning about P3P yet and we
 7   need to get the word out that this is not an expensive,
 8   time-consuming thing to do, particularly if you create a
 9   single policy statement in P3P.
10             Secondly, while it was easy to create and write
11   the code, companies do need better guidance for how to
12   implement P3P.  For example, we heard earlier that it does
13   make sense in some cases to have multiple P3P statements
14   for your web site.  That's counterintuitive to most
15   companies.  You think you want to have one statement, keep
16   it simply.  And by the way, that's more supportable as
17   well.  Companies need guidance on how to implement this in
18   a supportable way and also in a way, though, that's most
19   helpful for consumers.  And I think we're still learning
20   that.
21             As far as what else will spur companies to move
22   forward, certainly seeing a critical mass of consumers out
23   there with the capability to use P3P and evidence that
24   they're using it will spur.  What gets measured gets done. 
25   Similar to the kinds of measures that were publicized
0048
 1   about privacy statements in the past, we need to do the
 2   same thing with P3P.
 3             Secondly, organizations that are interested in
 4   moving this forward need to create sort of their own
 5   project management around this.  How many of the Fortune
 6   500 companies, how many of the top 100 web sites have done
 7   this?  If they haven't, call them up and find out why not.
 8             Thank you. 
 9             MR. MAXWELL:  Great.  
10             Ed, from a consumer standpoint do you want to
11   comment about what you've heard?
12             MR. MIERZWINSKI:  Thank you, Elliot.  The Public
13   Interest Research Group has views that represent those of
14   consumers, but more and more consumer protection has come
15   to take into account issues such as privacy and the
16   development of fair information practices.  Quite simply,
17   from our perspective we don't feel that the notice and
18   choice provisions enacted by the code of the machine are a
19   substitute for the full panoply of fair information
20   practices that includes so much more than notice and
21   choice that were originally embodied in the early 1970's
22   by the HEW task force that led to the inclusion in most of
23   the U.S. laws governing U.S. government uses of
24   information a set of practices that says, collect the
25   least information possible, collect it suitable for a
0049
 1   specific purpose, give consumers control over their
 2   information, not choice but consent-type control over
 3   their information, ensure that the information is
 4   protected by security standards, give the consumer the
 5   right to correct the information and to know all about the
 6   information in the database about him or her.
 7             I quite frankly don't see how P3P meets the fair
 8   information practices test.  The notion that notice is a
 9   privacy policy is patently absurd, and the notion that
10   notice and choice, what I call FIPS Lite, which is all
11   most industry groups want to give us, are adequate for
12   what the American public are clamoring for -- and this is
13   the American public that doesn't only include the members
14   of the so-called liberal groups such as the ACLU or the
15   Nader groups -- we're not really a Nader group, but
16   everybody thinks we are -- such as my group, but also
17   includes organizations such as Phyllis Schafly's
18   organization, the Eagle Forum, very conservative
19   organizations.  Senator Shelby is aligned with Congressman
20   Markey supporting bills that represent really strong fair
21   information practices.  So it's a broad section of the
22   public wants privacy protection.  
23             Ultimately, I think P3P will fail for a number
24   of reasons.  First of all, it's not really a negotiation. 
25   It's not a privacy-enhancing technology.  It's at best,
0050
 1   it's a -- I'm sorry.  At best it's a tool that allows the
 2   companies, as Mark Rotenberg has pointed out, it allows
 3   the web site to instantly know your privacy price. 
 4             What do you learn about them?  Virtually
 5   nothing.  Any consumer who wants to take advantage of very
 6   high P3P protocols is going to end up, is going to end up,
 7   I think, subject to all the nuisance screens that we get
 8   when we look at cookies.
 9             How many of you have set your cookie preference
10   to "Notify Me" or "Reject All Cookies" and then had
11   difficulty trying to surf?  You can't surf when you've got
12   those cookie pop-up windows coming up all the time, and so
13   people are going to give up on high privacy protection
14   under P3P and they're going to end up with low privacy
15   protection, and that's what the industry wants and that's
16   disappointing.
17             I think its supporters are going to say P3P INP,
18   P3P is no panacea, but it's something we can do now.  It's
19   not good enough and it's going to prevent the development
20   of better privacy protection policies that enhance
21   anonymity, and that's why we don't like it.
22             Thanks.
23             MR. MAXWELL:  Karen.
24             MS. COYLE:  Now we need this one on.
25             Notice is an important feature of any privacy
0051
 1   program and it is notice that is addressed by P3P. 
 2   However, as you just heard, notice does not by itself
 3   provide any amount of privacy.  With P3P it's like we now
 4   have the axle, but we are still lacking the wheels, the
 5   cart, and the horse.  We do not have a privacy solution.
 6             Proponents of P3P claim that the notice provided
 7   by web site privacy policies gives imminent users a
 8   choice.  This presumes that there will be comparable
 9   services that differ significantly only in their privacy
10   statements, and I see no indication that this will be the
11   case.
12             The invasion of privacy is deeply is entwined
13   with the reliance on advertising for revenue.  In a highly
14   competitive environment like the Internet today, the
15   winners are all under the same pressure to play the
16   customer profiling game.  In a world of information, what
17   is, after all, a comparable product?  If I want to read
18   the New York Times on line, but I do not want to give ut
19   information about myself to do so, reading another
20   newspaper that doesn't require me to sign up, say the San
21   Francisco Chronicle, definitely does not give me the same
22   content.
23             Unlike other products, information resources
24   tend to be unique.  As a matter of fact, that uniqueness
25   is encouraged by our copyright laws.  Where will the
0052
 1   reader turn for a choice?
 2             But even worse in this approach is that the
 3   approach places the burden on the Internet users to
 4   essentially shop for their own privacy.  I believe that
 5   privacy should be a right, not a bargain hunt.  I'm
 6   dismayed when P3P is touted as a solution.  It's only when
 7   we create the rest of the vehicle that we will actually
 8   enhance privacy on the Internet.  There are a number of
 9   commercial products that now address this issue.
10             But my hope is that we'll turn our attention to
11   the root of the problem and implement a baseline of
12   privacy that is the default for all users and in all
13   situations.  It's only then that privacy will be a right
14   and not a privilege enjoyed by the technologically elite
15   few.
16             Thank you. 
17             MR. MAXWELL:  Deirdre.
18             MS. MULLIGAN:  I hope you all aren't getting the
19   double echo that we are up here, because I've heard Karen
20   and Marc's presentation in like stereo from three
21   different directions.
22             I am in a position which is not as unusual as it
23   may seem from the panelists.  I'm a privacy advocate and I
24   fully support P3P.  I completely agree with people on both
25   sides of this issue if you could identify two of them, but
0053
 1   there actually aren't.  P3P is clearly not the silver
 2   bullet, but I don't think you've heard anybody up here
 3   suggest that it is.
 4             P3P is a very positive step in the direction of
 5   what Karen has called notice, Danny has called machine-
 6   readable access to information, and what I call
 7   transparency.  I can tell you, if I'm expected to shop
 8   with my feet and identify good privacy choices for myself
 9   and be an engaged consumer, certainly the ability to
10   access information in an easy fashion that doesn't burden
11   me with having to read the fine print -- I'm sure you've
12   all tried to look at the back of your Fair Credit
13   Reporting Act notice on your credit card and that's really
14   easy to do -- steps that promote transparency and that
15   enable consumers to diminish some of the costs of
16   protecting their privacy I think is a very positive step.
17             We've seen numerous surveys that say consumers
18   care, not a tiny little bit, but feverishly about their
19   privacy.  But it's very, very difficult for them to gain
20   access to the information needed to take steps to protect
21   it.  One of my favorite stories is Elliot Spitzer, who is
22   the attorney general in New York State, talked about three
23   of his senior staff attorneys, one of whom is a good
24   friend of mine, spending an awful lot of time trying to
25   decipher what a privacy statement meant at a major portal
0054
 1   site.
 2             I think if three very smart attorneys with a
 3   background in privacy can't understand a privacy policy,
 4   which is supposed to be at least the initial step in
 5   figuring out whether or not they want to exchange
 6   information with a business, that we have a real problem.
 7             Does P3P undermine privacy?  I personally don't
 8   think so.  I think we've seen more privacy activity at the
 9   state and federal level in the past two years than we have
10   in a very long time.  Most of our privacy laws are dated
11   from the 1970's.  We've seen increased pressure on the
12   private sector to develop standards.  There is certainly
13   an awful lot of effort needed to ensure those standards
14   are actually a race to the top, not a race to the bottom.
15             But I think that when we think about a little
16   sunlight as both disinfectant and also as motivation, that
17   transparency that P3P can bring is part of that sunlight. 
18   So I think that we have to continue to say that this is a
19   step forward.  It's clearly not the horse and the cart. 
20   But to suggest that it's a rock in the road and not an
21   axle I think really undermines all of our efforts to move
22   forward.
23             MR. MAXWELL:  Dan.
24             MR. JAYE:  As I'm here to speak somewhat from an
25   industry perspective, first of all I totally concur with
0055
 1   Deirdre that we have to look at progress towards a
 2   solution as not being inherently evil because it
 3   undermines the cause that we need to make bigger efforts
 4   to get to a final solution.  You never get to the final
 5   destination if you don't start making steps forward, and I
 6   think that, once again, everyone is in agreement that P3P
 7   is a mechanism for making it easier to understand and
 8   process notice and to enable tools that actually will make
 9   it easier to support the other fair information practices.
10             So it's a step forward, it's not a complete
11   solution.  But I think one of the most important aspects
12   of P3P is just the discipline it imposes upon the industry
13   and policymakers to codify privacy practices.  The P3P
14   vocabulary is enormously important because the process of
15   going through, creating a P3P policy statement, refines a
16   company's understanding of exactly what it is committing
17   to.
18             It's also turning out to be enablers of many
19   other things.  So for example there's another standards
20   effort going on called CP Exchange.  Actually, I have high
21   hopes that CP Exchange will be a mechanism for addressing
22   another major concern in privacy, which is onward
23   transfer.  A site may make many representations to a
24   consumer, but once that site transfers data to a third
25   party, even with the best intentions and contracts, the
0056
 1   horse is out of the barn.
 2             I think the ability of having additional
 3   standards that leverage the initial work of P3P to, for
 4   example, bind consumer data, encrypt it, and tie it with
 5   strong semantic information about what can and cannot be
 6   used with that information, such as you can use this
 7   information to ship a product, but you can't use it for
 8   anything else, will be very useful, but once again not
 9   complete solutions.
10             Once again, the final area that makes P3P very
11   important is that as a company that has tried to find the
12   balance between consumer privacy and the marketer's need
13   to be effective, to have an advertising-supported Internet
14   economy, is that we've tried very hard to follow solutions
15   that basically don't require us to need to know what an
16   individual is.  
17             You can market effectively without knowing who
18   an individual is, but just understanding certain
19   preferences.  You have to minimize data, you have to
20   follow specific practices.  Without a solution like P3P,
21   there's no ability for a company like ours to distinguish
22   ourselves from other companies who have very intrusive
23   data collection approaches.
24             MR. PRESLER-MARSHALL:  Thank you.  Is this one
25   on?
0057
 1             I'd like to point out that the World Wide Web is
 2   part of the real world and users have an expectation when
 3   they are dealing with an organization that they want to
 4   have trust in that organization.  That organization may be
 5   a corporation, that organization may be a government, that
 6   organization may be a nonprofit organization.  But in
 7   either case, individuals want to have trust in an
 8   organization that they're working with.
 9             As part of that process, they should expect to
10   know what an organization is going to do with their
11   information.  So if I come to a web site I want to know,
12   okay, what is this organization going to do with my
13   information.  For a web site, you've got to be part of
14   establishing that trust regime.
15             In any case, whatever kind of site it is that
16   you're running, you do need to be able to establish that
17   trust relationship so that you can see that, so that you
18   can see that person come back to you again and again.
19             P3P comes into this picture because P3P gives
20   users useful, actionable, understandable privacy
21   statements.  It lets people actually quickly and easily
22   understand what a site is going to do, what information
23   the site is going to make use of, and what choices the
24   user has in interacting with that site.
25             P3P is also very useful for the web site as
0058
 1   well.  It is useful in that, as Stan pointed out, by
 2   codifying your site's privacy statements you understand
 3   what it is that you're living up to.  In many
 4   organizations that can be complicated.  A large company
 5   has very many different branches or arms or portions of
 6   the company that may be interacting with users in
 7   different ways.  If a company is going to make a unified
 8   statement, P3P can help them understand, this is our
 9   statement, and then the applications that are making use
10   of information can be written to back up that statement.
11             When you put statements in machine-readable
12   format, you can process them automatically at a web site
13   as well or at a corporation as well, so that corporations
14   can really enact what they say they're interacting --
15   they're enacting, excuse me.
16             P3P is also useful in that it is implementable
17   and deployable at web sites at a reasonable cost.  As Mel
18   pointed out, it doesn't require that you reprogram your
19   web site, it doesn't require that you replace large
20   amounts of infrastructure at your web site.  P3P is
21   realizable and that's very important for a lot of web
22   sites.  I've heard tell that one web site managed to
23   deploy P3P in ten minutes based on an existing privacy
24   policy.  I was very impressed by that number.
25             Last of all, I want to point out that this is
0059
 1   P3P Version 1 and web protocols, networking protocols and
 2   software all will evolve based on the needs of their
 3   users.  In this case, those users are end users, those
 4   users are web sites and corporations.  This will move
 5   forward as it is needed, and I look forward to seeing
 6   what's going to happen with it in the future.
 7             MR. MAXWELL:  Thanks very much to all of you. 
 8   It's really heroic to try in this period of time just to
 9   try and sort through some of the basic issues.  I'd like
10   to open it up to members of the panel if they have
11   comments that they would like to make.
12             I have one question.  If we sort of stipulate
13   for the moment that it's not a solution and if I stipulate
14   for the moment that it's a tool and if we stipulate for
15   the moment -- and we don't have to believe this, but --
16   that everybody here is interested in increasing
17   individuals' control over their own information, one of
18   the issues -- I look at this audience and say, how many of
19   you have made adjustments to the settings of your browser? 
20   How many?
21             (A show of hands.)
22             So call it in this case about two-thirds of an
23   audience which is interested enough to come to the
24   Department of Commerce for a session on privacy
25   technology.  Now let's assume for the moment that we have
0060
 1   a much broader audience, 100 million people in the United
 2   States, 300 million people around the world.  What can we
 3   do to give people sufficient encouragement, tools,
 4   simplicity, ease of whatever, to make this tool useful, to
 5   make sure that people feel comfortable with it, that if
 6   they choose to approach it this way that they will say, I
 7   can do it and I can do it in a nanosecond and it's easy
 8   and it's just like buttering the toast in the morning? 
 9   How are we going to make it possible for people to really
10   make use of this tool if it does give people a greater
11   sense of empowerment?
12             MS. MULLIGAN:  I direct people to something
13   that's called the P3P Guiding Principles.  I think Elliot
14   is highlighting the importance, A, of educating consumers
15   about the existence of tools and the fact that they can
16   use them.  There are some efforts under way kind of in the
17   broader technology community, but also efforts like this,
18   to educate the public.
19             But it also really emphasizes the focus on how
20   does the product come out of the box, what does it look
21   like, what are the defaults, how configurable is it, how
22   obvious is it to the consumer?  And because this is what
23   we'd like to call, I think, a social protocol, it's not
24   just about technology, it's about a pressing social issue
25   -- privacy.
0061
 1             There was a lot of thought within the P3P
 2   working groups about how to give guidance to web sites, to
 3   implementers, about when they were designing products how
 4   they should think about designing them, make sure data
 5   isn't transferred unless a consumer explicitly wants to. 
 6   Now, P3P doesn't transfer data, but P3P might be built
 7   into a product that does.  It could be in a product that
 8   provides for anonymization, a product that reads P3P
 9   statements, and a product that provides a wallet.
10             Well, the guidance there is make sure that, even
11   though the P3P policy has been read, that the tool doesn't
12   automatically blast consumers' data away from them without
13   actually requiring some affirmative steps.  It's not a
14   one-click, it's a two-click.  If you look at the guiding
15   principles, there's a lot of direction.  Some of it's
16   should, some of it's must's, some of it's, we'd really
17   like you to.  But it's a lot of forward thinking about how
18   should the product come out of the box.
19             I think that there's going to need to be a lot
20   of vigilance from consumers and people what care about
21   privacy in evaluating products and make sure that they
22   actually meet the goals of advancing privacy.  But I think
23   it's a process and it's a very iterative one and we're
24   just in the beginning of that. 
25             MR. MAXWELL:  Marc, then Danny, then Karen.
0062
 1             MR. BEREJKA:  I would just like to second
 2   Deirdre's point about the importance of defaults and also
 3   reiterate my invitation to interested parties to work with
 4   Microsoft on our P3P implementation for the operating
 5   system.  We know that there are a lot of conflicting
 6   tensions, but one of the things that Microsoft does care
 7   about enormously is the user experience.  We bring people
 8   in off the street and we have them sit down and we have
 9   them hack around and we have them -- we get real consumer
10   feedback.
11             We also run public betas and try to get a lot of
12   feedback that way.  My ultimate point is that this is not
13   going to be an easy balancing act, but at least from our
14   perspective it's a balancing act that we look to
15   accomplish within the very real near-term.
16             MR. WEITZNER:  So you didn't ask people how many
17   buttered their toast in the morning.  That should have
18   been the question.
19             But I think that the real question should have
20   asked, or rather the question that you should ask to the
21   audience of web users out there, is how many of you set a
22   preference in AOL or whatever Internet service provider
23   that you use?  Granted, most people don't go and set their
24   browser preferences.  However, people develop expectations
25   about their browsers.
0063
 1             People, I think through a whole lot of public
 2   information effort and just common sense, are reluctant to
 3   enter credit card numbers when they don't see that little
 4   locker key icon closed.  That's not because they know what
 5   SSL is or because they went and changed some settings or
 6   because they downloaded more security.  It's because,
 7   through a pretty complex process, which we obviously all
 8   have to come to understand better, their expectations
 9   changed as a result of technology tools that were
10   available to protect them.
11             That's the kind of dynamic I think we're looking
12   for, and I think based on the experience with tools like
13   SSL we can expect to be successful.  I think it takes a
14   lot of effort, no question, and I don't think we should
15   just assume that people are going to become technological
16   geniuses.  But I think there is evidence that people
17   gradually do use new technology when it actually offers
18   them something.
19             MR. MAXWELL:  Karen.
20             MS. COYLE:  Well, I think there's even a prior
21   question that we need to ask here.  The Pew study which
22   Secretary Mineta alluded to basically says that a large
23   number of people are concerned about their privacy, but a
24   huge number of people, although a majority have heard
25   about cookies, they have no idea what it is.
0064
 1             I think the first question is who is going to
 2   educate the Internet users as to how it actually is that
 3   their privacy is being invaded, because it's only with
 4   that information that they will have the knowledge to turn
 5   to something like P3P.
 6             MR. MAXWELL:  Martin.
 7             MR. PRESLER-MARSHALL:  Elliot, I'll give you a
 8   direct answer to one of the questions you asked, which is
 9   how can we help see this get rolled out and then be usable
10   to individuals.  There are a wide variety of opinions
11   among the people in the world as a whole about what
12   constitutes an acceptable level of privacy.  There are a
13   large variety of opinions within my own household as to
14   what constitutes an acceptable level of privacy.
15             Part of this discussion is to give individuals
16   that flexibility.  There's a need for organizations to be
17   able to, for public interest organizations to be able to
18   express not necessarily defaults, but settings:  This is
19   what we believe is an acceptable level of privacy.  So if
20   you happen to be in the Ralph Nader camp, you may go to an
21   organization, a similar organization, and see, these are a
22   reasonable set of settings.  Or if you happen to be more
23   interested in getting highly personalized content on the
24   Internet, then maybe you can find different settings.
25             So this really needs to be a broad reach, with
0065
 1   many people offering opinions, because there are a wide
 2   variety of opinions within the United States and around
 3   the world.  Tim Burners-Levy, inventor of the World Wide
 4   Web, is fond of reminding people that the first "W" in
 5   "WWW" stands for "World," and there's a lot of opinions
 6   out there and we need to make sure that those are all
 7   supported.
 8             MR. MAXWELL:  Thank you all very much.  I'd like
 9   to be able to take questions from the audience if you have
10   them.  The only thing that you have to do is to respond to
11   the question about buttering.
12             If you'd identify yourself, please.
13             MS. WOODARD:  Yes, my name is Gwendolyn Woodard
14   and I would like to thank each of you for the information
15   on the tools that the consumer have to work with. 
16   However, I would like for you to talk about how these
17   tools will work on a voicing browser for individuals who
18   have physical challenges, and just could you talk about
19   that issue.
20             MR. WEITZNER:  Let's see.  It's an excellent
21   question and I'm not going to be able to give you a
22   complete answer.  You're probably familiar with the World
23   Wide Web Consortium's web accessibility initiative.  One
24   of the benefits of having machine-readable privacy
25   policies that I didn't mention is of course that once a
0066
 1   browser or a user agent has overall abilities to
 2   accommodate people's different disabilities, that browser
 3   can then present the information in the way that the user
 4   is comfortable with once it's encoded in machine-readable
 5   format.
 6             So the fact that, as Dan said, this is
 7   semantically structured information, the browser that
 8   knows to send that information through a braille reader or
 9   not to present it in image format but to do it in some
10   other format will be able to make those accommodations
11   much more effectively than a policy that's just written
12   out in English.
13             I'd also point out that the fact that these are
14   machine-readable policies means that they can be presented
15   in any number of natural languages that the user happens
16   to be comfortable with. 
17             MR. WEITZEL:  David Weitzel from Mitretek. 
18   We're a systems engineering nonprofit here in the suburbs. 
19   My question relates to a report or reports that came out
20   of GAO last week about the United States Government and
21   its interaction with the citizens.  Is there a role here
22   for the government to step up quickly to be an early
23   adopter and to lead the way as a good Internet citizen in
24   its interaction with the American populace?
25             MR. MAXWELL:  Rather than shunting that question
0067
 1   to any of the other people who might be willing on the
 2   panel to answer it:  Yes.  The Commerce Department is
 3   committed to having its pages P3P-compliant.  It's already
 4   well on its way to do that.  
 5             That report from the GAO I think was very
 6   troublesome, not because of what it concluded, but because
 7   of sort of what it asked and what it ignored.  What it
 8   ignored was something that Ed mentioned, which essentially
 9   is since the early seventies the government has been under
10   the restraints of the Privacy Act and we're all very
11   pleased that it does.  So the web sites that were queried,
12   were queried on the basis of a set of principles that were
13   not appropriate for the question.
14             So we feel, I think, in the administration quite
15   proud of the steps that have been taken to increase the
16   privacy protections of the citizen and to be able to work
17   with consumer groups, with the industry, to think about
18   how to increase individual control over information.  That
19   was I think a sort of fairly bum rap.  But sort of the
20   easy answer to the question is we are committed to P3P
21   implementation here and we're on our way.
22             Over here.
23             MR. CLARK:  Drew Clark with National Journal's
24   Technology Daily.
25             Mr. Jaye mentioned that P3P can help you compete
0068
 1   on privacy, on strong privacy protections.  But Mr.
 2   Mierzwinski and Ms. Coyle seem to say that there really
 3   won't be competition on privacy within certain industry
 4   spheres.  I'd like to understand exactly how privacy --
 5   how P3P will help you compete on privacy and to get the
 6   viewpoints of some of the other panelists on whether P3P
 7   will facilitate market competition on better privacy
 8   policies.
 9             MR. JAYE:  Thank you.  First of all, the ability
10   to conduct advertising on the Internet requires certain
11   mechanisms like cookies to be able to, for example, when
12   somebody clicks on an ad send them to the appropriate web
13   site, so we can basically staple together the resulting
14   page and the ad together, in Lorrie's analogy.  In
15   addition, advertisers want to know how many visitors saw
16   an ad.  So there are statistical purposes that are not
17   privacy invasive, that are not used to make decisions
18   about individuals, for which this information is used.
19             To be able to express that that's what we're
20   doing in certain cases and to be able to do that, because
21   we've been able to express it and distinguish ourselves
22   from unknown policies, is an enabling capability for an
23   advertising-supported model. 
24             But specifically with regard to privacy and
25   competitive advantage, major brands respond to consumers. 
0069
 1   Consumers care about privacy.  Major brands and
 2   advertisers don't want to be associated with bad actors. 
 3   Being able to maintain a position that is strong with
 4   regard to privacy has been a competitive advantage.  It's
 5   a position we've taken for five years, and it's our story
 6   and we're going to stick to it because it's working very
 7   well for us.
 8             MR. MAXWELL:  Because of time, let's just turn
 9   to the last question.
10             MR. STAMPLEY:  Dave Stampley from the New York
11   AG's Office.
12             For anybody who is of the mind that there should
13   be certain defaults and that the best thing to do for
14   consumers would be to recognize those defaults and hand
15   them to consumers, I guess my question is is there such a
16   thing as an identifiable default privacy value that in
17   fact is not a choice or value judgment itself that might
18   usurp some other consumer's power?   And might it be more
19   important instead to focus on where is a good baseline and
20   should all persons collecting information be obligated to
21   provide abilities to then vary from that baseline or let
22   consumers set their own preferences?
23             I'm just curious if there is a sense that there
24   is a way we could say, this is privacy and we know what it
25   looks like and we'll set it here at this point in time.
0070
 1             MS. MULLIGAN:  I think it's a great question and
 2   I think it's a place to distinguish between substantive
 3   defaults and process defaults.  What I mean is, my
 4   decisions about privacy and what would be an appropriate
 5   disclosure of information are going to vary depending upon
 6   the situation.
 7             For example, if I'm trying to get a driver's
 8   license on line, yes, they're covered by the Privacy Act. 
 9   B, I have no choice.  They're not going to give me the
10   driver's license unless I give them two forms of ID.  For
11   me in that situation, that's going to be acceptable
12   privacy.
13             Now, if CVS said that they're not going to let
14   me buy Rollos without two forms of ID, that's not going to
15   be acceptable.  That's a substantive privacy decision.
16             I think some of the process privacy decisions,
17   for example we were talking about does the information get
18   transferred without the consumer's affirmative action,
19   that same process I think you could apply in both
20   situations.  So even if I'm applying for a license, I
21   should have to actively hand over my two forms of ID. 
22   They shouldn't automatically be sucked out of my computer.
23             So I think there may be some areas, we may have
24   substantive defaults that we think apply in commercial
25   transactions, we may have substantive defaults that we
0071
 1   think apply in government transactions.  I think it would
 2   be very, very difficult and it's a whole other process to
 3   create those.
 4             I think Martin made an excellent point: 
 5   Consumers are going to feel differently because they have
 6   different concerns and they have had different
 7   experiences, and it's not going to be big broad cuts, you
 8   know, commerce, government.  it's going to be certain
 9   companies that I already do business with, or you're going
10   to have lots of variety.
11             But I do think on the process defaults that I
12   think that there is some progress to be made in thinking
13   about how to implement some of those in a broad way across
14   different kinds of implementations.
15             MR. PETERSON:  P3P is available, the client
16   tools at least in beta are available, to make consumer
17   research very possible.  So starting with what we know or
18   what we think we know and then putting it in the hands of
19   consumers worldwide, not just in the U.S., and seeing how
20   they react and whether it's useful or not is something we
21   ought to definitely be doing to really decide where these
22   things ought to net out.
23             MR. MAXWELL:  I'd like to thank the audience for
24   their questions.  I'd like to thank the panel.  It's
25   really a quite extraordinary group of people who I think
0072
 1   have worked very hard on this issue. 
 2             This privacy technology is only one piece of the
 3   puzzle.  It's clearly only one part, and when people sort
 4   of, I think, look at this and see it's sort of this or
 5   that, it's really not the kind of sophisticated analysis
 6   that you'd expect.  It's about law, it's about self-
 7   regulation, because self-regulation is just another word
 8   for what the companies will be doing for themselves and
 9   how they will think about it.  It's about consumer
10   awareness and education.  It's about how the technology
11   provides tools.
12             I think when we put all of these together, while
13   there will be differences, I think we all are committed to
14   exploring the issue of how can we give people more control
15   over information about themselves so that we can in fact
16   harvest the incredible technology that is available to us
17   now.  I think everybody working here makes a huge
18   contribution to that effort.
19             So thanks again for your time and effort and for
20   your attention. 
21             (Applause.)
22             MS. LEVY:  Thank you for your insights on P3P. 
23   Thanks for the panel.
24             I'd like to now invite everyone to go to a 15-
25   minute break.  I invite you out to the lobby to see the
0073
 1   exhibits and to enjoy some refreshments being hosted by
 2   the Internet Education Foundation.  Our next panel will
 3   begin at 11:15.
 4             (Recess from 11:01 a.m. to 11:28 a.m.)
 5             MS. LEVY:  Good morning.  We're going to get
 6   started with our second panel this morning.  This panel is
 7   on the role of privacy --
 8             VOICES:  Your mike's not on.
 9             MS. LEVY:  It's not on?  Can I try to get a mike
10   here?
11             (Pause.)
12               PANEL DISCUSSION:  IMPLICATIONS FOR
13              FAIR INFORMATION PRACTICE PRINCIPLES
14             Hello, can you hear me now?  Is that working?
15             We just want to start the second panel this
16   morning.  It's going to be on the role of privacy-
17   enhancing technologies and the fair information practice
18   principles.  We're pleased to have as our moderator for
19   this panel Dr. Lorrie Faith Cranor, what spoke this
20   morning and was introduced by Assistant Secretary Rohde. 
21   So I'm going to let Dr. Cranor go forward.  Thank you.
22             DR. CRANOR:  Thanks.  Let me start by
23   introducing the panelists and then we are going to -- I
24   think it's on.  We are going to go through a series of
25   questions and answers, rather than having the panelists
0074
 1   each give a presentation.  So let me go through the
 2   panelists.  They're all here now.
 3             First, we have Brian Adkins, who is Director of
 4   Government Relations for the Information Technology
 5   Industry Council, where he handles privacy, intellectual
 6   property, and other e-commerce issues.  He is also Co-
 7   Chairman of the Privacy Leadership Initiatives Technology
 8   Working Group.
 9             Next we have Scott Beechuk.  He is Co-Founder
10   and CEO of Privacy Right, Inc.  His previous technical
11   career in embedded systems design, object-oriented
12   programming, and engineering management served as the
13   basis for his deep interest in and understanding of
14   Internet privacy and security technology.
15             Next we have Glee Cady, who is Vice President of
16   Global Public Policy for Privada, Inc.  She brings over 20
17   years of technology and Internet experience as a respected
18   author, educator, technology executive, and policy
19   adviser.
20             To my left is Caitlin Halligan, who is Chief of
21   the New York Attorney General Elliot Spitzer's Internet
22   Bureau.  The Bureau coordinates statewide law enforcement
23   -- ooh, now I can really be heard -- statewide law
24   enforcement efforts regarding online consumer fraud,
25   privacy, securities trading, and other Internet-related
0075
 1   issues.
 2             Now I have an echo of myself in both ears.
 3             To my left we have Lance Hoffman, who is filling
 4   in for Joel Reidenberg, who had a family emergency.  Lance
 5   is Professor of Computer Science at the George Washington
 6   University.  He is in charge of the computer security
 7   graduate program in computer science.  he is the author or
 8   editor of five books and numerous articles on computer
 9   security and privacy, and he founded the School of
10   Engineering Cyberspace Policy Institute.  Lance and I both
11   served on the Advisory Committee on Online Access and
12   Security at the FTC.
13             Next we have Gary Laden.  Gary joined BBBOnline
14   on October 1st, 1998, as Director of the BBBOnline Privacy
15   Program.  From 1994 to September 1998 Gary served in the
16   Federal Communications Commission's Cable Services Bureau,
17   first as Chief of its Policy and Rules Division and most
18   recently as Chief of the Consumer Protection and
19   Competition Division.
20             Prior to his service at the FCC, Gary was at the
21   FTC for 21 years as an attorney and Assistant Director of
22   the Marketing Practices Division.
23             Next we have Stephanie Perrin, who is the Chief
24   Privacy Officer of ZeroKnowledge, formerly the Director of
25   Privacy Policy for Industry Canada's Electronic Commerce
0076
 1   Task Force.  Stephanie Perrin manages ZeroKnowledge's
 2   public affairs activities and acts as the company's
 3   primary liaison to government and nongovernmental
 4   organizations.  An internationally recognized expert in
 5   freedom of information and privacy issues, Stephanie was
 6   instrumental in developing Canada's privacy and
 7   cryptography policies over the past 15 years.
 8             Finally, we have Ari Schwartz, who is a policy
 9   analyst at the Center for Democracy and Technology.  Ari's
10   work focuses on protecting and building privacy
11   protections in the digital age by advocating for increased
12   individual control over personal information.  He also
13   works on expanding access to government information via
14   the Internet and online advocacy in civil society.
15             Ari is a leading expert on the issue of privacy
16   on government web sites and has testified before Congress
17   and Executive Branch agencies on the issue.
18             We're going to start today with Ari, who is
19   going to give us an overview of the fair information
20   practice principles so we're all on the same page and know
21   what we're talking about.
22             MR. SCHWARTZ:  I'm going to stand mostly because
23   I feel as though I'm facing that way and I'm speaking to
24   you.  Although I only have one slide, I could have 70, but
25   after Lorrie's experience I think I'm better off narrowing
0077
 1   it down to one. 
 2             (Screen.) 
 3             The Fair Information Practice standards are the
 4   basic standards by which we measure data privacy.  Many
 5   companies come to us and ask, what are the basic standards
 6   that we should address.  We hand them a list of Fair
 7   Information Practices.  The discussion doesn't end there,
 8   mostly because the standards -- the list of different
 9   kinds of standards cover different issues.  There are many
10   different areas, many different kinds of these standards,
11   and they're often portrayed in different lights.
12             Sometimes we've heard recently that a lot of
13   these Fair Information Practice standards, the sets that
14   we've been seeing, have been portrayed as international,
15   not American focused.  But in reality, the Fair
16   Information Practices are an American idea.  As was
17   mentioned on the last panel, in 1973 Health, Education and
18   Welfare Departments put together the first set of Fair
19   Information Practices.  Those were a set of four
20   practices, more like statements than actual bulleted ideas
21   as we see before us today.
22             Since that time we have seen these practices
23   grow and shrink in different formulations, usually ranging
24   between 4 and 12 different practices.  The most popular
25   that any self-respecting CPO, chief privacy officer, or
0078
 1   privacy analyst or policy analyst working on privacy
 2   should know by heart are:
 3             The Organization for Economic Cooperation and
 4   Development standards.  They have eight standards.  Those
 5   were made in 1980, agreed upon by all the OECD countries.
 6             The FTC standards.  That's the Federal Trade
 7   Commission standards, which have come out more recently. 
 8   They have a set of five standards, and sometimes these
 9   standards overlap, sometimes they don't.
10             These, the ones you see here, these are the
11   Department of Commerce's standards that they put forward. 
12   We see that they have chosen six.  The list that the
13   Center for Democracy and Technology covers chooses seven. 
14   If you want to see those, go to cdt.org.  That's my
15   commercial for the day.
16             But I've been asked to show the Department of
17   Commerce standards.  They're somewhat similar to the CDT
18   standards, and we will also hear from other people where
19   these are lacking, I'm sure, later in the panel.  But I
20   will try and do as good a job of covering them as I can.
21             Can people in the back read the full slide?
22             VOICES:  No.
23             MR. SCHWARTZ:  Then I will reread what it says
24   on the slide.  I'm sure that you can read some of the
25   bold.  The first practice listed is:  "Awareness. 
0079
 1   Companies should raise consumer awareness and should post
 2   privacy policies that articulate the manner in which it
 3   complies with other fair information practices."
 4             I've also noted here that this term is also
 5   called "notice" and many people on the panel will refer to
 6   it as notice, rather than awareness.  That's the way it's
 7   listed in the FTC principles, for example.
 8             That's the most basic of Fair Information
 9   Practices, just the idea that individuals should know
10   what's going on with personal information.
11             Second is what's called choice here:  "Companies
12   must give the opportunity to exercise choice with respect
13   to whether and how their personal information is used." 
14   In other words, that individuals just are given some kind
15   of choices.  Consent, opt-in, opt-out, are often terms
16   used when referring to this practice.  
17             Consent and opt-in are often used synonymously. 
18   Opt-in is the idea that individuals should be given the
19   ability to affirmatively consent to uses beyond the
20   transaction at hand.  So if the information is being used
21   for another purpose, the individual should be able to
22   consent for those uses.
23             Opt-out is when individuals -- when the
24   information is by default used for other purposes and the
25   individual is given the opportunity to get off of the
0080
 1   list.  There are many ways to frame opt-in, opt-in and
 2   opt-outs.
 3             Most of the debate that goes on today about
 4   choice are about whether it should be opt-in or opt-out. 
 5   I'd like to also put out the idea that there's a third way
 6   here, which is not just opt-in or opt-out, but give the
 7   individual two choices -- I want my information shared, I
 8   don't want my information shared -- right up front,
 9   getting beyond this kind of idea that there is some kind
10   of default that needs to be set.
11             Third on this list is security:  "Companies must
12   take reasonable precautions to protect data from loss,
13   misuse, alteration, or destruction."  This is a pretty
14   straightforward principle that's in almost every set of
15   fair information practices.
16             Data integrity:  "Companies should only collect
17   and keep personal data relevant for the purposes for which
18   it has been gathered.  The data should be accurate,
19   complete, and current."  This is covering two or three
20   different types of practices:  one, the first set, which
21   some consider to be collection limitation, and there are
22   different ways that that can be framed, the basic idea
23   that you should only be collecting information relevant
24   for the purpose; and that the data should be accurate and
25   complete, is also in most fair information practices and
0081
 1   often called data integrity.
 2             Access:  "Companies should offer consumers
 3   reasonable access to information about them and a means to
 4   correct or amend inaccurate information."  This is an oft
 5   hotly debated subject.  This was, as Lorrie mentioned,
 6   this was covered recently in an Access and Security
 7   Working Group that the Federal Trade Commission held. 
 8   They put out a very good report on this subject and on
 9   security, which both Lorrie and Lance were on, I think.  I
10   don't know if anyone else on this panel was, but some of
11   the other panelists have been as well.
12             That really goes through each of those issues
13   about where people lie on this.  It ranges from people
14   should just know some basic information held about them,
15   what has been collected in the past, to that individuals
16   should have access and be able to correct anything that is
17   held about them, and there's different places in between
18   there, which I'm sure will be addressed.
19             Accountability:  "Companies must be accountable
20   for complying with their privacy policies."  This one is
21   also known as enforcement and it is also somewhat basic
22   and straightforward.
23             DR. CRANOR:  Thank you. 
24             Stephanie, I was wondering if you would care to
25   comment on some of the other principles that are not part
0082
 1   of this set that you think we should be thinking of as
 2   well.
 3             MS. PERRIN:  Thanks, Lorrie.  I guess, should I
 4   speak here or go up to the podium?
 5             DR. CRANOR:  From the seat.
 6             MS. PERRIN:  From the seat, okay.  I think
 7   Lorrie's afraid I'll get up there and talk for half an
 8   hour.
 9             In Canada, about ten years ago, recognizing the
10   difficulties that we were facing in this area and frankly
11   the lack of compliance to the OECD guidelines, we formed a
12   committee under the Canadian Standards Association, which
13   is a recognized standards development body, and created
14   basically a management standard for privacy based on the
15   OECD guidelines with a view to making what are good
16   statements of principle auditable.
17             We came up with a list of ten.  Just briefly, we
18   pulled some of them forward and gave them the kind of
19   depth that is impossible to get in an international body
20   such as the OECD when you're working on these things, and
21   we put what we thought was the first one forward, and that
22   is accountability:  Every organization shall be
23   accountable for the information under its care and shall
24   put in place procedures and practices to give effect to
25   that accountability, shall educate, shall make people
0083
 1   aware, shall name someone within the organization
 2   accountable.  It's a pretty full principle.
 3             I should just add that Canada went ahead once we
 4   had agreed on this voluntary standard, which took
 5   basically about four years -- we had an industry, a
 6   consumer rep, and government committee, 47 members.  Some
 7   of you have some idea what that would be like.  And we
 8   pounded this out.  Once it became a standard, we move
 9   ahead and legislated.
10             So if you're interested in looking at it, it is
11   the Schedule 1 of the recently passed Personal Information
12   Protection and Electronic Documents Act, which cleared
13   Parliament in April and will come into force in Canada for
14   all organizations engaged in commercial activity on
15   January 1.
16             So we pulled the accountability up forward.  We
17   gave a lot more emphasis to the pieces that are necessary
18   in management systems to make it auditable and make people
19   accountable.
20             The next one was purpose specification.  You had
21   to state your purpose.  There were huge fights over
22   whether you could -- whether we would get into this
23   standard the concept of whether a purpose was legitimate. 
24   Obviously, businesses weren't keen on that, consumers
25   were.  So we at least said that you had to state it.  I
0084
 1   think that complies partly with your awareness, only it is
 2   more fuller in this standard.
 3             Consent.  One of my problems with the opt-in,
 4   opt-out is it's a totally opt-in, opt-out.  You opt in for
 5   something, you're there for the ride.  Really, with the
 6   complexity of the personal information flows in the data
 7   age, we've got to be able to articulate that consent on a
 8   data element basis.  
 9             There are certain provisions under this consent
10   clause.  One that I would pull to your attention is you
11   are not required -- and this is now law, of course -- to
12   give more information than is required for the delivery of
13   a particular product or service, and if the company denies
14   you the product or service based on a failure to give
15   information not necessary for it, you've got a justified
16   complaint.
17             That kind of thing doesn't come across in a
18   simple opt-in, opt-out.  So a little more articulation on
19   the consent.
20             There would be limits to collection.  We see
21   different thresholds here.  Instead of putting collection,
22   use, and disclosure all in one bailiwick, there's
23   collection tied to purpose and then use and disclosure
24   tied to purpose.  So they're in two separate principles. 
25   You limit the collection first and then once you've got
0085
 1   that data you limit the use, the disclosure, and the
 2   retention of that data, because as long as data's hanging
 3   around of course the temptation to use it for another
 4   purpose arises.
 5             Accuracy, a little different.  This insists on
 6   accuracy, this data integrity principle.  In our debates
 7   we discovered that, frankly, it wasn't in a consumer's
 8   interest to insist on accurate information because it gave
 9   companies an excuse for a fishing trip to go back and get
10   recent, more accurate data.  If you don't need it, you
11   don't need to make it more accurate.  In other words, if
12   you've got a loan ten years old, don't go back and look
13   for my new income if I'm making my payments.
14             So that's a little different.  We have a
15   safeguards principle that I think is a little fuller,
16   although that's a good security principle there. 
17   Basically, industry standards is what we're looking for,
18   and that's of course where our companies come in, is to
19   make this real.
20             Openness, similar to your first principle, but
21   basically a company has to make policies, procedures, down
22   to the detailed collection instrument level available to
23   people on request.  That doesn't mean you have to have it
24   all there in your office, but if an individual inquires
25   that stuff has to be made available.  That's a little
0086
 1   fuller, I think, than just an education sort of imperative
 2   that you see in that first principle.  Again, under the
 3   law, of course, people can complain on any of these
 4   things.
 5             Individual access, hotly debated, of course, in
 6   Canada, as it is everywhere.  But it's a fundamental right
 7   of privacy.  If you don't have the right to get your
 8   records and to change them and to have whatever your view
 9   of the story is travel with the records wherever they go,
10   then you don't really have a human rights-based approach
11   to this in our view.  So that's pretty strong under this
12   bill.
13             Challenge.  Many data protection statutes and
14   indeed the OECD guidelines gave people a right to
15   challenge the accuracy of their own information and some
16   of them the use that's being put to it.  We broadened that
17   because we recognize that a lot of the problems in privacy
18   arose over security issues, over the way the company
19   handles the data. 
20             So under this standard and now piece of
21   legislation, you have the right to challenge any of the
22   practices, and of course in the regime we have you can
23   take that to a privacy commissioner and have it
24   investigated.
25             Now, this standard has been put forward for
0087
 1   several years to ISO as the basis for a management
 2   standard.  It's had, I think, three or four unanimous
 3   resolutions of the COPOCO committee -- that's the Consumer
 4   Policy Committee of ISO -- endorsing it, but it keeps
 5   getting blocked.  So the ad hoc advisory group that was
 6   looking at this for a potential for an international
 7   standard has now been disbanded.  There's work going on in
 8   Europe looking at standards, but I think that's where it
 9   ends.
10             DR. CRANOR:  Thanks.
11             Now that we have the overview of what the
12   principles are, let's focus on technology, which is the
13   focus of this panel.  I'd like to turn to Glee and ask you
14   to describe how the technology that your company offers
15   can support some of these principles.
16             MS. CADY:  Good morning.  Is this on?  Ah, good. 
17   I'm now getting the echo that Lorrie was talking about
18   earlier.
19             I find it very difficult to answer this question
20   because we don't see our technology as other than parallel
21   to the Fair Information Practices.  What we're trying to
22   do is to build an infrastructure where you don't have to
23   count on the other party being good.  Do we support
24   philosophically all of these things that Ms. Perrin so
25   eloquently described?  The answer to that is yes.
0088
 1             But we also support basically the idea that,
 2   because in an Internet environment what we have is a
 3   bottom-up structure of making it relatively easy for new
 4   companies to come online all the time and making it
 5   relatively difficult for us to get the word out, us
 6   collectively as those people being interested in the
 7   privacy community, about what is responsible behavior,
 8   what we're afraid of is that the trust that comes that
 9   enables good things to happen, whether you're sharing
10   information or you're buying products, won't be based on
11   the necessary experience that we find in the real world.
12             So what we're trying to do at Privada is to
13   build privacy into the infrastructure.  That's somewhat
14   hubristic at this point, all right.  The current product
15   sets that we offer and are offered by other anonymizing
16   and pseudonymizing companies don't provide a total
17   package.  We're all working toward that, but again this is
18   a bottom-up process where we're working in conjunction
19   with top-down standards like P3P in order to provide the
20   accurate notice, so that you could tell what someone's
21   privacy policies are if you visit their site, but also
22   that you don't have to count on it because you don't have
23   to tell them who you are in order to achieve what you need
24   from them.
25             So our motto is "Privacy Under Your Control." 
0089
 1   We encourage responsible behavior on the part of
 2   individuals certainly and on the part of other companies,
 3   but we're trying to make it so the consumer doesn't have
 4   to know the four principles of so on and the five
 5   principles of someone else or the six principles of the
 6   Department of Commerce in order to be both protected and
 7   effective.
 8             DR. CRANOR:  Thanks.
 9             Caitlin, I'd like you to talk about what you see
10   as the limitations of technology in addressing these
11   principles.
12             MS. HALLIGAN:  I think there are a couple of
13   different ways -- the echo is disturbing, right.  There
14   are a couple of different ways of getting at -- no echo.
15             There are a couple of different ways of getting
16   at that question.  I think all of the technologies we've
17   been looking at today only address discrete components of
18   the Fair Information Practices, and I don't think any of
19   them purport to address all of them.  So if we take P3P,
20   for example, it I think does take some important steps
21   towards improving notice and improving consumer ability to
22   understand the information practices that a site might
23   have.
24             By doing that, it also facilitates choice.  It
25   doesn't, I think, do as much with respect to enhancing
0090
 1   access and security.  I think that's okay.  I think,
 2   secondly, if you look at how these technologies function
 3   to protect consumers with respect to particular aspects of
 4   Fair Information Practices that they try to get at, we
 5   could all identify limitations and I think some of those
 6   are a function of where the technology is today and where
 7   it might be tomorrow.
 8             For example, again if you look at P3P, one of
 9   the pieces that doesn't seem to be on the table right now
10   in this current iteration is the ability for consumers to
11   negotiate on a real-time basis with a site with which they
12   might want to interact.  So if a site's privacy policies
13   don't match their preferences, there's not an opportunity
14   to offer to engage in some kind of trade.  But again
15   that's a function, one would hope, of where the technology
16   is today.
17             I think the third way of looking at this
18   question and maybe the most important is whether
19   technology, whatever it's able to do for us, is sufficient
20   to fulfil Fair Information Practices.  I think the answer
21   to that is probably not.  I think that we do need some
22   sort of statutory guideline that puts rules in place and
23   creates incentives for the technologies to develop in a
24   way that promotes those principles.  I think that's true
25   for a couple of reasons.
0091
 1             First of all, these technologies, as wonderful
 2   as they might be, are not self-adopting.  They're not
 3   self-adopting for businesses and they're not self-adopting
 4   for consumers, either.  For consumers to take advantage,
 5   for example, of Privada they have to understand that there
 6   are issues out there about information practices that
 7   might trouble them and that there is an option out there
 8   for them to protect their privacy.
 9             Also, I think that these technologies don't
10   address and don't again purport to address a second very
11   important component here, which is Stephanie's first
12   principle, and that's accountability.  There aren't a lot
13   of ways in which the enforcement is really enhanced by
14   these technologies.  There aren't ways in which there are
15   audits readily done by these technologies which allow for
16   -- whether it's a regulatory entity or watchdog
17   organization, to easily monitor whether in fact sites are
18   in compliance with the practices that they set forth to
19   the public.
20             So I think that there is a bigger picture
21   question out there that's worth thinking about.
22             DR. CRANOR:  Before we go on, does anyone want
23   to disagree with any of those limitations or point out any
24   technologies that we haven't maybe thought of? 
25             (No response.) 
0092
 1             DR. CRANOR:  Okay.  Then what I want to do now
 2   is -- oh, go ahead.
 3             MS. PERRIN:  I hate to grab the mike again,
 4   Lorrie, because I did have quite a bit of time there.  But
 5   I think we have to be clear.  I agree with our last
 6   speaker, a technology is not a replacement for a law.  It
 7   is a heck of a good way to implement the requirements of a
 8   law and, frankly, that's why I came to ZeroKnowledge,
 9   because the next challenge is to build the principles of
10   law, the principles of what we agree on here, into the
11   infrastructure.
12             I think that many of these things become web
13   enforceable w