1
00:00:00,000 --> 00:00:19,450


2
00:00:19,450 --> 00:00:20,250
Welcome back.

3
00:00:20,250 --> 00:00:26,530


4
00:00:26,530 --> 00:00:29,820
Our next panel, or our
first panel, actually,

5
00:00:29,820 --> 00:00:32,729
deals with an interesting
and not so widely discussed

6
00:00:32,729 --> 00:00:36,390
technology along
with its associated

7
00:00:36,390 --> 00:00:37,965
crucial social question.

8
00:00:37,965 --> 00:00:41,060


9
00:00:41,060 --> 00:00:43,330
Danny Weitzner, a
lawyer by training,

10
00:00:43,330 --> 00:00:46,690
a pioneer of the negotiated
approach to privacy,

11
00:00:46,690 --> 00:00:50,350
and head of the Laboratories
Worldwide Web Consortium

12
00:00:50,350 --> 00:00:53,830
Technology and Society
Division will chair the panel

13
00:00:53,830 --> 00:00:56,060
and frame the question.

14
00:00:56,060 --> 00:00:57,440
Danny?

15
00:00:57,440 --> 00:01:00,058
Thank you, Bob.

16
00:01:00,058 --> 00:01:01,850
[LAUGHTER]

17
00:01:01,850 --> 00:01:06,380
See, I knew that he
would steal my opening.

18
00:01:06,380 --> 00:01:09,080
We're here to talk about a
really important subject,

19
00:01:09,080 --> 00:01:12,230
pranks notwithstanding.

20
00:01:12,230 --> 00:01:14,964
I think we'd all agree that--

21
00:01:14,964 --> 00:01:17,946
[LAUGHTER]

22
00:01:17,946 --> 00:01:18,940


23
00:01:18,940 --> 00:01:22,630
I think that we'd all agree that
one of the tremendous paradoxes

24
00:01:22,630 --> 00:01:24,920
of the net is that
on the one hand,

25
00:01:24,920 --> 00:01:27,900
our interactions online
seem so impersonal,

26
00:01:27,900 --> 00:01:30,400
and this frustrates
people sometimes

27
00:01:30,400 --> 00:01:33,230
and leads for calls for
greater trust online.

28
00:01:33,230 --> 00:01:36,580
But at the same time, many of
us are aware that we really

29
00:01:36,580 --> 00:01:38,770
do leave a trail of
virtual breadcrumbs

30
00:01:38,770 --> 00:01:43,180
as we surf around the web,
and this compels some of us

31
00:01:43,180 --> 00:01:47,590
to take rather extraordinary
measures to hide our identity,

32
00:01:47,590 --> 00:01:51,610
to conceal our true
identity when we're online.

33
00:01:51,610 --> 00:01:55,480
And one of these tools
is an anonymous remailer,

34
00:01:55,480 --> 00:01:57,910
one such remailer
is run here at LCS

35
00:01:57,910 --> 00:02:01,720
is the subject of a significant
amount of research activity.

36
00:02:01,720 --> 00:02:04,240
And Michael Dertouzos,
the head of the lab,

37
00:02:04,240 --> 00:02:08,020
has posed a very pointed
question to all of us,

38
00:02:08,020 --> 00:02:10,090
should the lab be
running this tool?

39
00:02:10,090 --> 00:02:13,090
Is this a valid
subject for research?

40
00:02:13,090 --> 00:02:15,580
Is this a socially
useful activity?

41
00:02:15,580 --> 00:02:17,530
Or is it destructive?

42
00:02:17,530 --> 00:02:20,950
And to answer these
questions, which

43
00:02:20,950 --> 00:02:23,770
are very important, especially
important to some people

44
00:02:23,770 --> 00:02:27,370
on the panel, we have a
really tremendous panel

45
00:02:27,370 --> 00:02:31,420
of experts who are going
to help us sort out

46
00:02:31,420 --> 00:02:35,290
both the core legal and
technology questions posed

47
00:02:35,290 --> 00:02:37,840
by anonymous remailers.

48
00:02:37,840 --> 00:02:39,400
At the end of the
table on the left

49
00:02:39,400 --> 00:02:42,910
is Nadine Strossen, who's
a personal hero of mine,

50
00:02:42,910 --> 00:02:45,820
the president of
the ACLU since 1991

51
00:02:45,820 --> 00:02:47,770
and a professor at
New York Law School.

52
00:02:47,770 --> 00:02:49,663
She's been a national leader.

53
00:02:49,663 --> 00:02:51,430
[APPLAUSE]

54
00:02:51,430 --> 00:02:51,930
Thank you.

55
00:02:51,930 --> 00:02:54,205
Now, was that for New York
Law School or the ACLU?

56
00:02:54,205 --> 00:02:57,710


57
00:02:57,710 --> 00:03:00,230
She has been a national leader
in defending civil liberties

58
00:03:00,230 --> 00:03:01,250
on the internet.

59
00:03:01,250 --> 00:03:04,310
Her book, Defending
Pornography, Free Speech, Sex,

60
00:03:04,310 --> 00:03:08,630
and the Fight for Women's Rights
was named by The New York Times

61
00:03:08,630 --> 00:03:11,360
Notable Book of 1995.

62
00:03:11,360 --> 00:03:14,300
Next, we have Philip
Reitinger, a colleague

63
00:03:14,300 --> 00:03:18,140
at a truly honorable
adversary, sometimes, of mine,

64
00:03:18,140 --> 00:03:19,820
in these internet
civil liberties

65
00:03:19,820 --> 00:03:21,840
debates in Washington.

66
00:03:21,840 --> 00:03:23,900
He is senior counsel
with the United States

67
00:03:23,900 --> 00:03:26,630
Department of Justice
Computer Crimes section.

68
00:03:26,630 --> 00:03:30,350
He's prosecuted a number
of computer crime cases

69
00:03:30,350 --> 00:03:32,030
and is really a
leading authority

70
00:03:32,030 --> 00:03:35,250
on this subject in the country.

71
00:03:35,250 --> 00:03:40,550
We have some guy sitting here,
and next to him we have David

72
00:03:40,550 --> 00:03:43,520
Maziéres, who's a graduate
student here at LCS.

73
00:03:43,520 --> 00:03:46,130
He's doing research and
operating systems, security,

74
00:03:46,130 --> 00:03:47,480
and privacy.

75
00:03:47,480 --> 00:03:51,410
He designed and implemented
the nym.alias.net pseudonymous

76
00:03:51,410 --> 00:03:54,320
remailer, which is
currently running at LCS.

77
00:03:54,320 --> 00:03:57,980
We're going to try to have
a relatively open free form

78
00:03:57,980 --> 00:03:59,900
discussion among the panelists.

79
00:03:59,900 --> 00:04:02,510
To start us off,
David has agreed

80
00:04:02,510 --> 00:04:04,970
to spend about 10 minutes
just bringing us all up

81
00:04:04,970 --> 00:04:07,208
to speed on just what
this technology is

82
00:04:07,208 --> 00:04:08,250
that we're talking about.

83
00:04:08,250 --> 00:04:11,126
So David, please.

84
00:04:11,126 --> 00:04:14,060
[APPLAUSE]

85
00:04:14,060 --> 00:04:16,510


86
00:04:16,510 --> 00:04:17,529
Thanks.

87
00:04:17,529 --> 00:04:21,790
There are many reasons you might
want to send anonymous mail.

88
00:04:21,790 --> 00:04:23,440
So here's an example.

89
00:04:23,440 --> 00:04:24,903
This is a piece of email.

90
00:04:24,903 --> 00:04:26,320
And the person who
sent this email

91
00:04:26,320 --> 00:04:27,940
might not want to be traced.

92
00:04:27,940 --> 00:04:29,560
Now, you can look
at this and see

93
00:04:29,560 --> 00:04:33,430
that the message is addressed
to Bill Gates at Microsoft.com

94
00:04:33,430 --> 00:04:38,745
and it's coming from
pieguy@nym.alias.net.

95
00:04:38,745 --> 00:04:41,230
Now, what is nym.alias.net?

96
00:04:41,230 --> 00:04:44,680
Well, nym.alias.net
is a pseudonym server

97
00:04:44,680 --> 00:04:46,360
that we've been running at LCS.

98
00:04:46,360 --> 00:04:49,780
And this allows people to
establish email pseudonyms

99
00:04:49,780 --> 00:04:53,050
that we call nyms, and that look
like ordinary email addresses,

100
00:04:53,050 --> 00:04:56,410
like, for example,
pieguy@nym.alias.net.

101
00:04:56,410 --> 00:04:59,170
And nyms function just like
regular email addresses,

102
00:04:59,170 --> 00:05:02,770
and that you can use one both
to send and receive mail.

103
00:05:02,770 --> 00:05:04,900
But what's
interesting about nyms

104
00:05:04,900 --> 00:05:07,300
is that nym owners are
completely anonymous.

105
00:05:07,300 --> 00:05:10,120
So even the people who
administer the system

106
00:05:10,120 --> 00:05:12,460
don't know the true
identities of the people who

107
00:05:12,460 --> 00:05:15,100
own these nym email addresses.

108
00:05:15,100 --> 00:05:18,520
So nym.alias.net is actually in
fairly widespread use outside

109
00:05:18,520 --> 00:05:20,140
of computer science research.

110
00:05:20,140 --> 00:05:23,710
It's been open to the
public since June of '96,

111
00:05:23,710 --> 00:05:26,590
and you only need
a PGP to use it.

112
00:05:26,590 --> 00:05:30,520
PGP is a widely available
piece of encryption software.

113
00:05:30,520 --> 00:05:34,150
And people actually have gone
ahead and developed more user

114
00:05:34,150 --> 00:05:37,480
friendly clients specifically
for the server software that

115
00:05:37,480 --> 00:05:39,730
runs on DOS windows and Unix.

116
00:05:39,730 --> 00:05:41,920
And throughout the
existence of the service,

117
00:05:41,920 --> 00:05:45,910
there have been about 2,000
to 3,000 active accounts on it

118
00:05:45,910 --> 00:05:48,510
at any given time.

119
00:05:48,510 --> 00:05:51,410
Now, why do people need a
service like nym.alias.net?

120
00:05:51,410 --> 00:05:53,578
Well, of course, we
don't know, since we

121
00:05:53,578 --> 00:05:56,120
don't read people's email, we
have no way of knowing for sure

122
00:05:56,120 --> 00:05:57,980
what people are using
the service for.

123
00:05:57,980 --> 00:06:01,137
But we decided to send out
a survey to all of our users

124
00:06:01,137 --> 00:06:03,470
and ask them to tell us what
they like about the service

125
00:06:03,470 --> 00:06:04,790
and what they're using it for.

126
00:06:04,790 --> 00:06:08,000
And we encouraged people to
be as candid as possible,

127
00:06:08,000 --> 00:06:10,280
saying that this was not a
referendum on the service's

128
00:06:10,280 --> 00:06:12,470
existence, but would
help us in our research.

129
00:06:12,470 --> 00:06:15,110
And we got back a wide
range of responses

130
00:06:15,110 --> 00:06:19,040
that vary in the degree of
privacy that they needed.

131
00:06:19,040 --> 00:06:20,900
At the most critical
end of the spectrum

132
00:06:20,900 --> 00:06:23,120
were people who
use nym.alias.net

133
00:06:23,120 --> 00:06:25,820
because they need protection
from oppressive governments.

134
00:06:25,820 --> 00:06:27,470
One particularly
compelling story

135
00:06:27,470 --> 00:06:30,140
came from a person who
was a humanitarian aid

136
00:06:30,140 --> 00:06:32,840
worker in a country with a
fairly oppressive government.

137
00:06:32,840 --> 00:06:35,030
And that person said
that he or she did not

138
00:06:35,030 --> 00:06:39,050
feel comfortable communicating
back home to friends via email

139
00:06:39,050 --> 00:06:42,540
without the availability
of a service like this.

140
00:06:42,540 --> 00:06:43,922
Another large
class of people use

141
00:06:43,922 --> 00:06:45,380
the service because
they're worried

142
00:06:45,380 --> 00:06:49,100
about losing their jobs,
or potentially harassment,

143
00:06:49,100 --> 00:06:52,530
or just simply
embarrassment for uses,

144
00:06:52,530 --> 00:06:55,340
such as discussing
alcoholism, depression, being

145
00:06:55,340 --> 00:06:56,600
a sexual minority.

146
00:06:56,600 --> 00:06:59,660
Some people said that they
used nym.alias.net to blow

147
00:06:59,660 --> 00:07:02,180
the whistle on
illegal activities,

148
00:07:02,180 --> 00:07:06,020
and a few people said that
they found it incredibly

149
00:07:06,020 --> 00:07:08,510
useful for fighting
harmful cults

150
00:07:08,510 --> 00:07:10,790
and helping people
escape those cults.

151
00:07:10,790 --> 00:07:14,180
Another class of people simply
wanted protection from mail

152
00:07:14,180 --> 00:07:14,810
logs.

153
00:07:14,810 --> 00:07:18,290
Right now, on most systems,
if you have an email account,

154
00:07:18,290 --> 00:07:20,660
even if no one's sitting
there reading your mail,

155
00:07:20,660 --> 00:07:23,427
the system is building up
a log file of all the mail

156
00:07:23,427 --> 00:07:25,010
messages that come
and go, and someone

157
00:07:25,010 --> 00:07:27,440
can derive from that list
of all the people you've

158
00:07:27,440 --> 00:07:28,910
exchanged email with.

159
00:07:28,910 --> 00:07:30,920
And people who don't feel
comfortable with that

160
00:07:30,920 --> 00:07:32,840
use this service to
keep the identity

161
00:07:32,840 --> 00:07:36,620
of their correspondence secret
from the system administrators.

162
00:07:36,620 --> 00:07:40,580
And finally, some people simply
use this service for protection

163
00:07:40,580 --> 00:07:41,630
from search engines.

164
00:07:41,630 --> 00:07:43,370
And the fact is
right now you could

165
00:07:43,370 --> 00:07:46,130
make a statement in a public
forum under your real name,

166
00:07:46,130 --> 00:07:50,090
and 10 years from now people may
call up your name on a search

167
00:07:50,090 --> 00:07:51,710
engine and be
looking at statements

168
00:07:51,710 --> 00:07:52,700
you made 10 years ago.

169
00:07:52,700 --> 00:07:54,830
Do you still want
to be associated

170
00:07:54,830 --> 00:07:55,790
with those statements?

171
00:07:55,790 --> 00:07:59,420
And in particular, one person
gave a story of a candidate

172
00:07:59,420 --> 00:08:03,272
at a job interview who's Usenet
News posting history came up

173
00:08:03,272 --> 00:08:04,730
as a subject for
the job interview,

174
00:08:04,730 --> 00:08:07,970
and that person was obviously
very uncomfortable with that.

175
00:08:07,970 --> 00:08:10,250
Finally, a few people
admitted to using

176
00:08:10,250 --> 00:08:12,938
the service for admittedly
marginal purposes.

177
00:08:12,938 --> 00:08:14,480
And probably the
most common of these

178
00:08:14,480 --> 00:08:16,340
are people who said
they use it to discuss

179
00:08:16,340 --> 00:08:18,170
marijuana cultivation.

180
00:08:18,170 --> 00:08:19,970
But there are also
a couple of people

181
00:08:19,970 --> 00:08:22,220
who said that they used
it for software piracy

182
00:08:22,220 --> 00:08:26,330
and for virus development.

183
00:08:26,330 --> 00:08:28,960
So how do we build
a system like this

184
00:08:28,960 --> 00:08:32,980
that allows people to have these
email addresses anonymously?

185
00:08:32,980 --> 00:08:35,830
Well, the key idea is
that users of the system

186
00:08:35,830 --> 00:08:39,340
never communicate directly
with nym.alias.net.

187
00:08:39,340 --> 00:08:42,520
And the way we do this is
to build on an existing

188
00:08:42,520 --> 00:08:44,350
network of anonymous remailers.

189
00:08:44,350 --> 00:08:47,230
And these are servers
that forward email on

190
00:08:47,230 --> 00:08:51,550
to some destination address, but
strip off all the information

191
00:08:51,550 --> 00:08:54,400
about the person who
originated the piece of email.

192
00:08:54,400 --> 00:08:58,150
And there exist independently
operated in several countries

193
00:08:58,150 --> 00:08:59,480
and jurisdictions.

194
00:08:59,480 --> 00:09:03,070
And when mail is sent through
a series of remailers,

195
00:09:03,070 --> 00:09:05,050
the key point is that
multiple retailers

196
00:09:05,050 --> 00:09:07,000
need to be compromised
in order to expose

197
00:09:07,000 --> 00:09:09,880
the path that any email
message has taken.

198
00:09:09,880 --> 00:09:12,760
And so you can see
from this that it's

199
00:09:12,760 --> 00:09:15,100
fairly simple for
users to send mail

200
00:09:15,100 --> 00:09:18,910
to nym.alias.net when they want
to establish accounts and send

201
00:09:18,910 --> 00:09:20,020
outgoing mail.

202
00:09:20,020 --> 00:09:21,550
And there are
actually several ways

203
00:09:21,550 --> 00:09:24,550
of receiving anonymous
mail for nym.alias.net

204
00:09:24,550 --> 00:09:27,010
to reply to the users.

205
00:09:27,010 --> 00:09:29,260
And I'll give you a particularly
simple example, which

206
00:09:29,260 --> 00:09:31,900
is that even if you
forget about email,

207
00:09:31,900 --> 00:09:34,360
if you wanted someone to send
you a message without knowing

208
00:09:34,360 --> 00:09:36,220
who you are, you
might tell that person

209
00:09:36,220 --> 00:09:38,672
to go take out a classified
ad in The New York Times,

210
00:09:38,672 --> 00:09:40,630
and you could read The
New York Times every day

211
00:09:40,630 --> 00:09:42,130
along with millions
of other people,

212
00:09:42,130 --> 00:09:44,050
and see the message
when it arrives.

213
00:09:44,050 --> 00:09:48,700
And nym.alias.net allows people
to configure their pseudonyms

214
00:09:48,700 --> 00:09:51,772
so that any mail that is
received for that pseudonym

215
00:09:51,772 --> 00:09:53,980
is posted to a public forum
where it's very difficult

216
00:09:53,980 --> 00:09:55,300
to trace to the person.

217
00:09:55,300 --> 00:09:59,320
And there are also ways of doing
this more efficiently using

218
00:09:59,320 --> 00:10:02,000
the anonymous remailer network.

219
00:10:02,000 --> 00:10:04,190
OK, so the first
thing that we noticed

220
00:10:04,190 --> 00:10:06,980
when we deployed the system
is that nym.alias.net

221
00:10:06,980 --> 00:10:08,330
draws attacks.

222
00:10:08,330 --> 00:10:11,000
Anonymous speech can
definitely upset people.

223
00:10:11,000 --> 00:10:13,273
And when people get upset
at anonymous speakers,

224
00:10:13,273 --> 00:10:14,690
the first thing
they want to do is

225
00:10:14,690 --> 00:10:17,000
know who is speaking
anonymously so that they can

226
00:10:17,000 --> 00:10:18,980
exact some kind of retribution.

227
00:10:18,980 --> 00:10:21,375
And failing that, when
they can't find out

228
00:10:21,375 --> 00:10:23,000
who's actually making
these statements,

229
00:10:23,000 --> 00:10:24,920
they tend to turn on
the service provider

230
00:10:24,920 --> 00:10:26,480
and they want to shut it down.

231
00:10:26,480 --> 00:10:30,320
And in fact, protecting
the identities of users

232
00:10:30,320 --> 00:10:32,330
is a fairly well-known problem.

233
00:10:32,330 --> 00:10:34,970
And it's been
going back to 1981,

234
00:10:34,970 --> 00:10:36,380
there's been work in that.

235
00:10:36,380 --> 00:10:38,570
And we understand how
to do this fairly well.

236
00:10:38,570 --> 00:10:41,870
But in practice, defending
a service and keeping it up

237
00:10:41,870 --> 00:10:44,060
and useful is actually not
nearly as well understood

238
00:10:44,060 --> 00:10:44,907
a problem.

239
00:10:44,907 --> 00:10:47,240
And so I'll give you some
examples of some of the things

240
00:10:47,240 --> 00:10:50,370
you can do to try to disrupt
an anonymous service.

241
00:10:50,370 --> 00:10:52,520
One thing you can do
is use the service

242
00:10:52,520 --> 00:10:54,710
to attack some third party
anonymously, and then

243
00:10:54,710 --> 00:10:57,140
hope that that third party
turns on the remailer

244
00:10:57,140 --> 00:10:58,890
and tries to get it shut down.

245
00:10:58,890 --> 00:11:01,580
Another thing you
can try to do is

246
00:11:01,580 --> 00:11:05,420
to try to marginalize the
service through disruptive use

247
00:11:05,420 --> 00:11:09,950
so that if the
existence of the service

248
00:11:09,950 --> 00:11:12,710
becomes absolutely
intolerable to everyone,

249
00:11:12,710 --> 00:11:15,442
then even legitimate
users of the system

250
00:11:15,442 --> 00:11:17,900
are not going to be able to be
heard because no one's going

251
00:11:17,900 --> 00:11:20,780
to have any choice but to ignore
all anonymous traffic that's

252
00:11:20,780 --> 00:11:22,100
coming from the service.

253
00:11:22,100 --> 00:11:24,140
And finally, one thing
you can try to do

254
00:11:24,140 --> 00:11:26,750
is to try to make
life really difficult

255
00:11:26,750 --> 00:11:28,310
for the operators
of the service,

256
00:11:28,310 --> 00:11:30,310
so that it will not be
worth their while to keep

257
00:11:30,310 --> 00:11:33,470
running the service and
they'll want to shut it down.

258
00:11:33,470 --> 00:11:35,930
Well, in fact, defending
against some of these attacks

259
00:11:35,930 --> 00:11:37,940
also poses some
unusual challenges.

260
00:11:37,940 --> 00:11:41,280
And there are several
reasons for that.

261
00:11:41,280 --> 00:11:43,880
First of all, in general,
when you have a server that's

262
00:11:43,880 --> 00:11:46,550
on the internet and you're
trying to protect it

263
00:11:46,550 --> 00:11:48,525
from attack, one of your
front line defenses,

264
00:11:48,525 --> 00:11:50,900
and the best ways of finding
out if you're being attacked

265
00:11:50,900 --> 00:11:52,370
and potentially
who's attacking you

266
00:11:52,370 --> 00:11:55,250
is to look at logs of who's
connecting to the server

267
00:11:55,250 --> 00:11:55,880
and using it.

268
00:11:55,880 --> 00:11:58,160
And in general, with
an anonymous server,

269
00:11:58,160 --> 00:12:00,840
privacy concerns preclude
any kind of usage logs.

270
00:12:00,840 --> 00:12:04,130
You don't want to keep track of
who's made use of your service.

271
00:12:04,130 --> 00:12:06,950
A second problem is that the
service by its very nature

272
00:12:06,950 --> 00:12:09,680
is designed to hide the
identities of users.

273
00:12:09,680 --> 00:12:11,480
And that applies
to abusers, too.

274
00:12:11,480 --> 00:12:13,970
So when someone is
abusing the system,

275
00:12:13,970 --> 00:12:16,400
you have no way of knowing
who that person is.

276
00:12:16,400 --> 00:12:19,190
In fact, even if you
do know who is actually

277
00:12:19,190 --> 00:12:21,530
trying to disrupt
use of your service,

278
00:12:21,530 --> 00:12:23,990
you can't necessarily
ban that person

279
00:12:23,990 --> 00:12:27,080
from making use of the service
because the server is designed

280
00:12:27,080 --> 00:12:30,680
to prevent the operators from
associating incoming messages

281
00:12:30,680 --> 00:12:32,240
with particular
people, you're not

282
00:12:32,240 --> 00:12:34,550
going to know which incoming
requests are actually

283
00:12:34,550 --> 00:12:38,480
coming from a malicious person
trying to disrupt your service.

284
00:12:38,480 --> 00:12:40,700
And finally, for a
number of reasons,

285
00:12:40,700 --> 00:12:43,040
any kind of
content-based censorship

286
00:12:43,040 --> 00:12:45,350
of filtering
disruptive messages is

287
00:12:45,350 --> 00:12:47,760
completely impractical
for a number of reasons.

288
00:12:47,760 --> 00:12:50,540
First of all, if you try to put
some kind of automatic filter

289
00:12:50,540 --> 00:12:53,120
in place that's just going
to filter on keywords,

290
00:12:53,120 --> 00:12:55,530
well, you're dealing
with human adversaries

291
00:12:55,530 --> 00:12:57,530
and they're going to adapt
to whatever automatic

292
00:12:57,530 --> 00:13:00,710
filters you put in place if
they know what you're blocking.

293
00:13:00,710 --> 00:13:04,130
Second of all, if you start
filtering certain messages,

294
00:13:04,130 --> 00:13:07,160
you're going to block legitimate
users, too, potentially.

295
00:13:07,160 --> 00:13:10,400
And in fact, this may well
be the goal of your attacker.

296
00:13:10,400 --> 00:13:14,240
So if your response to attacking
is to start filtering messages,

297
00:13:14,240 --> 00:13:15,865
and people want you
to filter messages,

298
00:13:15,865 --> 00:13:17,282
they're going to
attack the system

299
00:13:17,282 --> 00:13:19,170
and try to make
disruptive use of it.

300
00:13:19,170 --> 00:13:20,750
Finally, if you
actually try to look

301
00:13:20,750 --> 00:13:23,390
at every individual message
going through the system,

302
00:13:23,390 --> 00:13:25,730
it's just simply way
too much human effort.

303
00:13:25,730 --> 00:13:30,110
And in fact, in the US at least,
inspecting messages manually

304
00:13:30,110 --> 00:13:32,138
like that actually opens
you up to liability,

305
00:13:32,138 --> 00:13:33,680
which is something
you certainly want

306
00:13:33,680 --> 00:13:36,620
to avoid in the case of
an anonymous remailer.

307
00:13:36,620 --> 00:13:40,010
So some things we've concluded
from running this service.

308
00:13:40,010 --> 00:13:42,680
First of all, we think that
open anonymous servers, in fact,

309
00:13:42,680 --> 00:13:43,913
are practical.

310
00:13:43,913 --> 00:13:45,830
We've been running this
system for three years

311
00:13:45,830 --> 00:13:48,390
and we've had some hard
challenges to deal with.

312
00:13:48,390 --> 00:13:50,920
But basically, we've overcome
all the problems that

313
00:13:50,920 --> 00:13:52,320
have been thrown in our way.

314
00:13:52,320 --> 00:13:54,110
But one thing we have
learned is that when

315
00:13:54,110 --> 00:13:56,068
you're building an
anonymous service like that,

316
00:13:56,068 --> 00:13:59,850
you have to factor abuse into
the design of the service.

317
00:13:59,850 --> 00:14:01,970
A second thing we've
learned is that people

318
00:14:01,970 --> 00:14:05,510
can be incredibly
creative in finding ways

319
00:14:05,510 --> 00:14:07,520
to disrupt an anonymous server.

320
00:14:07,520 --> 00:14:10,190
So you really ought to
deploy anonymous systems

321
00:14:10,190 --> 00:14:12,390
to assess their viability.

322
00:14:12,390 --> 00:14:15,200
And in fact, when you're trying
to assess their viability,

323
00:14:15,200 --> 00:14:17,750
the key metric to
use in measuring

324
00:14:17,750 --> 00:14:20,570
how successful and robust
an anonymous system is,

325
00:14:20,570 --> 00:14:21,920
is really human time.

326
00:14:21,920 --> 00:14:23,840
And the worst thing
you can possibly

327
00:14:23,840 --> 00:14:26,030
do if you're designing
an anonymous system

328
00:14:26,030 --> 00:14:28,670
and deploying it is to put
yourself in a situation

329
00:14:28,670 --> 00:14:31,490
where manually you have to
inspect every message that goes

330
00:14:31,490 --> 00:14:34,430
through the system, or you have
to manually clean up the mess

331
00:14:34,430 --> 00:14:36,260
every time someone
manages to make

332
00:14:36,260 --> 00:14:38,720
disruptive use of the system,
because then someone can

333
00:14:38,720 --> 00:14:41,228
write a program to generate
all kinds of disruptive mail.

334
00:14:41,228 --> 00:14:42,770
And that person runs
a program that's

335
00:14:42,770 --> 00:14:45,350
costing you countless
hours every time it runs.

336
00:14:45,350 --> 00:14:48,080
And on the flip
side, the way to win,

337
00:14:48,080 --> 00:14:50,750
and the way to control abuse,
and to keep the service useful

338
00:14:50,750 --> 00:14:53,810
for legitimate users
is to try to limit

339
00:14:53,810 --> 00:14:56,480
the trouble to any
single person can stir up

340
00:14:56,480 --> 00:14:58,430
by making sure that
the system costs

341
00:14:58,430 --> 00:15:00,620
a certain amount of
human time to use.

342
00:15:00,620 --> 00:15:02,480
And I haven't been
able to go into any

343
00:15:02,480 --> 00:15:04,813
of the details of the kinds
of attacks that we had here,

344
00:15:04,813 --> 00:15:06,230
but if this is
interesting to you,

345
00:15:06,230 --> 00:15:08,480
I'd invite you to read some
of our war stories, which

346
00:15:08,480 --> 00:15:11,550
are on the LCS web page
under human impact.

347
00:15:11,550 --> 00:15:12,864
Thank you.

348
00:15:12,864 --> 00:15:16,343
[APPLAUSE]

349
00:15:16,343 --> 00:15:20,330


350
00:15:20,330 --> 00:15:22,560
Thank you, David.

351
00:15:22,560 --> 00:15:25,220
So Phil Reitinger,
can you give us

352
00:15:25,220 --> 00:15:27,140
a little bit of context here?

353
00:15:27,140 --> 00:15:31,790
At some level, this sounds like
pretty interesting research.

354
00:15:31,790 --> 00:15:35,315
David and this other
guy have worked hard.

355
00:15:35,315 --> 00:15:36,440
How do you know it's a guy?

356
00:15:36,440 --> 00:15:38,410
[LAUGHS] I was
sitting next to him.

357
00:15:38,410 --> 00:15:42,800


358
00:15:42,800 --> 00:15:45,470
What are we losing here?

359
00:15:45,470 --> 00:15:46,290
What's the problem?

360
00:15:46,290 --> 00:15:47,750
That calls for a
follow-up comment.

361
00:15:47,750 --> 00:15:48,250
Sorry.

362
00:15:48,250 --> 00:15:50,750


363
00:15:50,750 --> 00:15:52,880
I'd like to answer
your question but I'm

364
00:15:52,880 --> 00:15:56,500
frozen into fear by the
anonymous masked person

365
00:15:56,500 --> 00:15:57,188
next to me.

366
00:15:57,188 --> 00:15:59,640
[LAUGHTER]

367
00:15:59,640 --> 00:16:00,140


368
00:16:00,140 --> 00:16:01,040
Sure.

369
00:16:01,040 --> 00:16:02,780
You're not going to
unmask him, are you?

370
00:16:02,780 --> 00:16:05,420
Well, the upshot of that
is, of course, anonymity

371
00:16:05,420 --> 00:16:07,940
has a downside, too.

372
00:16:07,940 --> 00:16:10,220
I know this will
come as a shock,

373
00:16:10,220 --> 00:16:13,940
to many of you, including
Danny, but people actually

374
00:16:13,940 --> 00:16:16,350
commit crimes on the internet.

375
00:16:16,350 --> 00:16:20,930
And it's hard for
law enforcement

376
00:16:20,930 --> 00:16:24,350
to put a pseudonym in jail.

377
00:16:24,350 --> 00:16:26,250
You don't see a lot of
indictments that say

378
00:16:26,250 --> 00:16:28,880
United States versus John Doe.

379
00:16:28,880 --> 00:16:33,080
If you want to discover people
who are committing crimes,

380
00:16:33,080 --> 00:16:37,010
and then prosecute them, and put
them in jail, if appropriate,

381
00:16:37,010 --> 00:16:39,170
you've got to
discover who they are.

382
00:16:39,170 --> 00:16:43,220
In fact, law enforcement
has to prove identity

383
00:16:43,220 --> 00:16:45,350
in every criminal case.

384
00:16:45,350 --> 00:16:49,440
But anonymity is the
antithesis of identity.

385
00:16:49,440 --> 00:16:53,220
And so if a person
is truly anonymous,

386
00:16:53,220 --> 00:16:55,820
then there can be no
criminal law enforcement

387
00:16:55,820 --> 00:17:01,310
and people are free, which
has an upside and a downside,

388
00:17:01,310 --> 00:17:04,609
to doing whatever they
want on the internet.

389
00:17:04,609 --> 00:17:07,190
Now, one of the
things I said was

390
00:17:07,190 --> 00:17:10,698
that if a person is
completely anonymous-- it's

391
00:17:10,698 --> 00:17:12,740
an important distinction,
because a lot of people

392
00:17:12,740 --> 00:17:16,550
think about anonymity as
being a binary concept,

393
00:17:16,550 --> 00:17:19,250
either you are anonymous
or you are not.

394
00:17:19,250 --> 00:17:21,410
But in point of fact,
there's a fairly broad set

395
00:17:21,410 --> 00:17:24,740
of differentiation
depending upon how difficult

396
00:17:24,740 --> 00:17:29,480
it is to determine the real
identity of a person associated

397
00:17:29,480 --> 00:17:33,570
with a particular message
or a pseudonymous identity.

398
00:17:33,570 --> 00:17:36,500
And that's where the
distinctions come in.

399
00:17:36,500 --> 00:17:41,300
If one is serious about
enforcing law regarding

400
00:17:41,300 --> 00:17:44,900
activities on the internet--
and by enforcing law,

401
00:17:44,900 --> 00:17:46,910
I'm talking about
laws that refer

402
00:17:46,910 --> 00:17:50,150
to both speech-related
crimes, if you will,

403
00:17:50,150 --> 00:17:56,840
such as harassment, extortion,
and such similar sorts

404
00:17:56,840 --> 00:17:59,750
of crimes, and non
speech-related crimes

405
00:17:59,750 --> 00:18:02,210
like engaging in
computer intrusions,

406
00:18:02,210 --> 00:18:05,210
law enforcement has to
be able to determine

407
00:18:05,210 --> 00:18:07,670
the identity of the person.

408
00:18:07,670 --> 00:18:11,300
And so you need the ability
to trace communications

409
00:18:11,300 --> 00:18:13,640
to their source to determine
the identity of the person

410
00:18:13,640 --> 00:18:16,380
committing the crime.

411
00:18:16,380 --> 00:18:19,660
Let me just ask you, I
think that we understand

412
00:18:19,660 --> 00:18:22,350
the problem in the abstract.

413
00:18:22,350 --> 00:18:25,080
In point of fact, have
any of your investigations

414
00:18:25,080 --> 00:18:27,810
or investigations you're
aware of been blocked or made

415
00:18:27,810 --> 00:18:32,840
substantially more difficult
by anonymous remailers?

416
00:18:32,840 --> 00:18:37,850
I know of cases where
anonymous remailers were used

417
00:18:37,850 --> 00:18:40,830
in the commission of crimes.

418
00:18:40,830 --> 00:18:46,280
And I think anecdotally
there is a large upswing

419
00:18:46,280 --> 00:18:50,810
in the use of anonymous,
particularly web-based email

420
00:18:50,810 --> 00:18:54,350
services, to send
harassing communications.

421
00:18:54,350 --> 00:18:56,720
A typical example
might be, let's say

422
00:18:56,720 --> 00:18:59,690
you wanted to threaten the
President of the United States.

423
00:18:59,690 --> 00:19:02,762
So you go to an anonymous
web browsing service--

424
00:19:02,762 --> 00:19:04,970
don't want to mention any
of them-- but you go to one

425
00:19:04,970 --> 00:19:08,360
and you use that service to
access an anonymous email

426
00:19:08,360 --> 00:19:11,030
account and send out a harassing
message to the president

427
00:19:11,030 --> 00:19:13,550
or whatever person you
want, because it's likely

428
00:19:13,550 --> 00:19:17,840
that neither the anonymous
remailer or the anonymous web

429
00:19:17,840 --> 00:19:20,840
browsing service is going
to keep any records at all

430
00:19:20,840 --> 00:19:23,300
that would allow tracing
of that communications.

431
00:19:23,300 --> 00:19:26,180
You have effectively
committed the perfect crime.

432
00:19:26,180 --> 00:19:28,850
It's an almost
untraceable threat.

433
00:19:28,850 --> 00:19:31,335
And for example, if you
make it fairly detailed

434
00:19:31,335 --> 00:19:32,960
and make it a bomb
threat or something,

435
00:19:32,960 --> 00:19:36,950
you can cause a lot of
damage without really

436
00:19:36,950 --> 00:19:39,800
facing any real jeopardy.

437
00:19:39,800 --> 00:19:43,820
Nadine, Phil's very worried
about these services

438
00:19:43,820 --> 00:19:46,160
and he hasn't told
us what exactly he

439
00:19:46,160 --> 00:19:48,350
wants to do-- we'll
come to that later

440
00:19:48,350 --> 00:19:51,270
about these sorts of services.

441
00:19:51,270 --> 00:19:54,500
What do you think are the
issues on the free speech

442
00:19:54,500 --> 00:19:56,360
side of this equation?

443
00:19:56,360 --> 00:19:58,820
Aren't people just hiding
behind these barriers

444
00:19:58,820 --> 00:20:00,860
for no particularly good reason?

445
00:20:00,860 --> 00:20:04,130
Well, I, like Phil,
am very concerned

446
00:20:04,130 --> 00:20:06,120
about law enforcement.

447
00:20:06,120 --> 00:20:12,860
And I, like Phil, recognize that
the supreme law in this country

448
00:20:12,860 --> 00:20:15,440
is the United
States Constitution.

449
00:20:15,440 --> 00:20:18,470
I want to be sure that
constitutional guarantees

450
00:20:18,470 --> 00:20:21,420
of freedom of speech,
freedom of the press,

451
00:20:21,420 --> 00:20:24,410
freedom of association,
and privacy are

452
00:20:24,410 --> 00:20:27,830
as honored and
robust in cyberspace

453
00:20:27,830 --> 00:20:31,190
as they have been in
more traditional media.

454
00:20:31,190 --> 00:20:33,770
It proves too much
to say, well, we

455
00:20:33,770 --> 00:20:38,210
might be able to solve more
crimes, or detect more crimes,

456
00:20:38,210 --> 00:20:42,050
or even deter more
crimes if we somehow

457
00:20:42,050 --> 00:20:47,030
bit that supreme law to
alter the balance of power

458
00:20:47,030 --> 00:20:50,660
between individual freedom on
the one hand and government

459
00:20:50,660 --> 00:20:53,760
surveillance capability
on the other hand.

460
00:20:53,760 --> 00:20:57,380
The reason it proves too
much is that very same

461
00:20:57,380 --> 00:21:00,770
philosophical issue
confronted the framers.

462
00:21:00,770 --> 00:21:05,570
So I am taking the very
conservative position adopted

463
00:21:05,570 --> 00:21:10,820
200 years ago that there are
certain individual rights that

464
00:21:10,820 --> 00:21:16,160
simply may not be sacrificed,
no matter what the potential law

465
00:21:16,160 --> 00:21:18,900
enforcement gain is
on the other side.

466
00:21:18,900 --> 00:21:22,490
Another very, very important
point is we cannot let

467
00:21:22,490 --> 00:21:28,010
the debate be driven by thinking
of the most horrible abuses

468
00:21:28,010 --> 00:21:31,670
that could be made of
freedom of speech or privacy,

469
00:21:31,670 --> 00:21:35,540
because after all,
anything can be abused.

470
00:21:35,540 --> 00:21:38,600
When the automobile
was first invented,

471
00:21:38,600 --> 00:21:41,840
Phil's predecessors in the
Justice Department and law

472
00:21:41,840 --> 00:21:46,610
enforcement argued that it
should be an illegal object,

473
00:21:46,610 --> 00:21:48,620
because, after
all, criminals were

474
00:21:48,620 --> 00:21:51,350
going to be able
to use cars to flee

475
00:21:51,350 --> 00:21:53,070
from the scene of the crime.

476
00:21:53,070 --> 00:21:55,580
Now, that seems like an
absurd argument to us

477
00:21:55,580 --> 00:21:57,530
now, judging by the
audience response,

478
00:21:57,530 --> 00:22:00,680
but to me is exactly
the same argument that

479
00:22:00,680 --> 00:22:05,330
is being made now, because
some technology can be abused,

480
00:22:05,330 --> 00:22:07,790
can be used to commit
crimes, therefore, we

481
00:22:07,790 --> 00:22:10,760
are going to deprive all
of the legitimate users

482
00:22:10,760 --> 00:22:15,620
from human rights dissidents,
to whistleblowers, to authors,

483
00:22:15,620 --> 00:22:20,240
to all kinds of individuals
who, in many cases,

484
00:22:20,240 --> 00:22:24,410
would not have freedom
of speech if they did not

485
00:22:24,410 --> 00:22:28,070
have the security and
confidentiality provided

486
00:22:28,070 --> 00:22:30,170
by anonymous remailers.

487
00:22:30,170 --> 00:22:32,760
Now, I think at this
juncture, the guy in the mask

488
00:22:32,760 --> 00:22:34,010
might have something to offer.

489
00:22:34,010 --> 00:22:35,690
But he's going to
have to take it off

490
00:22:35,690 --> 00:22:37,325
before he can speak
to this audience.

491
00:22:37,325 --> 00:22:39,980


492
00:22:39,980 --> 00:22:44,300
This is a professor Frans
Kaashoek, who many of you know.

493
00:22:44,300 --> 00:22:46,363
He's associate
professor at MIT'S

494
00:22:46,363 --> 00:22:48,530
Electrical Engineering
computer. science Department.

495
00:22:48,530 --> 00:22:52,490
And he's a member of LCS.

496
00:22:52,490 --> 00:22:55,280
Aside from his work on
anonymous remailers,

497
00:22:55,280 --> 00:22:57,650
his research
includes exokernels,

498
00:22:57,650 --> 00:22:59,750
and extensible operating
system architecture,

499
00:22:59,750 --> 00:23:04,100
and SFSS, a secure decentralized
global file system.

500
00:23:04,100 --> 00:23:07,220
Frans, why did you
choose to do this work?

501
00:23:07,220 --> 00:23:10,410


502
00:23:10,410 --> 00:23:11,680
Good question.

503
00:23:11,680 --> 00:23:16,170
I guess I like getting
subpoenas and receiving

504
00:23:16,170 --> 00:23:18,360
the blame for this project.

505
00:23:18,360 --> 00:23:19,500
I always like to joke that.

506
00:23:19,500 --> 00:23:23,400
Dave gets the credit,
I get the blame.

507
00:23:23,400 --> 00:23:26,590
I guess it comes down
to a couple of issues.

508
00:23:26,590 --> 00:23:31,290
First, I guess I strongly
believe in freedom of speech.

509
00:23:31,290 --> 00:23:36,630
And I think this is going to be
very important issue in trying

510
00:23:36,630 --> 00:23:39,900
to figure out how to
have freedom of speech

511
00:23:39,900 --> 00:23:41,040
on the internet.

512
00:23:41,040 --> 00:23:44,370
And understanding that
and providing services

513
00:23:44,370 --> 00:23:46,620
that actually allow you
to have freedom of speech

514
00:23:46,620 --> 00:23:50,310
seems like crucial in
understanding the debate

515
00:23:50,310 --> 00:23:52,170
and the issues that
are involved here.

516
00:23:52,170 --> 00:23:55,350
And I guess go back
to, and sort of one

517
00:23:55,350 --> 00:24:00,780
of the comments Nadine makes
that the original federal

518
00:24:00,780 --> 00:24:05,040
papers were which were
United States was founded on

519
00:24:05,040 --> 00:24:06,780
were published on a pseudonym.

520
00:24:06,780 --> 00:24:11,340
And so there's clearly
a case that where

521
00:24:11,340 --> 00:24:16,370
anonymity is actually important
and plays an important role.

522
00:24:16,370 --> 00:24:19,670
Now, we might not
have been sitting here

523
00:24:19,670 --> 00:24:23,100
having been that the
case 200 years ago.

524
00:24:23,100 --> 00:24:26,150
The other side of
it is it brings out

525
00:24:26,150 --> 00:24:29,630
a lot of interesting
research problems

526
00:24:29,630 --> 00:24:31,460
in the area of security.

527
00:24:31,460 --> 00:24:35,030
I know Dave touched on a
number of them in his talk.

528
00:24:35,030 --> 00:24:39,140
Services like this
attract people

529
00:24:39,140 --> 00:24:40,640
that actually get
mad at the service

530
00:24:40,640 --> 00:24:42,920
and trying to shut
down the service.

531
00:24:42,920 --> 00:24:46,450
nym.alias.net is not the only
service that people are mad at.

532
00:24:46,450 --> 00:24:48,260
There are lots of
people out there

533
00:24:48,260 --> 00:24:50,420
trying to attack other
kinds of services

534
00:24:50,420 --> 00:24:52,650
and trying to force them down.

535
00:24:52,650 --> 00:24:55,400
So it seems to be
from a research point,

536
00:24:55,400 --> 00:24:58,820
important to try to understand
what kind techniques can

537
00:24:58,820 --> 00:25:02,870
be used to protect
against people

538
00:25:02,870 --> 00:25:05,510
that illegally are
shutting down services.

539
00:25:05,510 --> 00:25:09,170
I'm sure you have been reading
the attacks on the NATO website

540
00:25:09,170 --> 00:25:11,040
in the last couple of weeks.

541
00:25:11,040 --> 00:25:13,370
What kind of
techniques can develop

542
00:25:13,370 --> 00:25:18,440
to protect against these
denial of service attacks?

543
00:25:18,440 --> 00:25:22,520
And nym.alias.net is
one example of a service

544
00:25:22,520 --> 00:25:26,600
that we can investigate, learn,
understand, and figure out what

545
00:25:26,600 --> 00:25:29,670
techniques work and don't work.

546
00:25:29,670 --> 00:25:32,930
And part of that is, if you
offer a service like that,

547
00:25:32,930 --> 00:25:36,680
I think one of the key problems
is, how do you deal with abuse?

548
00:25:36,680 --> 00:25:38,863
It's not only focusing
on the problem

549
00:25:38,863 --> 00:25:40,280
of how do you
protect the identity

550
00:25:40,280 --> 00:25:43,610
of the nym, the
other side of that

551
00:25:43,610 --> 00:25:46,800
is, how do you deal with abuse?

552
00:25:46,800 --> 00:25:49,730
Phil gave his example
of harassment.

553
00:25:49,730 --> 00:25:52,970
It turns out that actually
is pretty easy to deal with.

554
00:25:52,970 --> 00:25:53,880
Two things.

555
00:25:53,880 --> 00:25:56,690
First of all, a piece of email
coming from nym.alias.net

556
00:25:56,690 --> 00:25:58,970
is identifying that
it came from there.

557
00:25:58,970 --> 00:26:02,210
Second of all, we have
features on the servers that

558
00:26:02,210 --> 00:26:05,570
allow users to say, I don't
want to receive an email coming

559
00:26:05,570 --> 00:26:06,855
from nym.alias.net.

560
00:26:06,855 --> 00:26:08,480
So one of the things
we're trying to do

561
00:26:08,480 --> 00:26:11,660
is to figure out how to make
the servers in such a way

562
00:26:11,660 --> 00:26:19,490
that people can live pleasantly
with the service in existence.

563
00:26:19,490 --> 00:26:23,458
And I guess, then,
the final side of this

564
00:26:23,458 --> 00:26:25,000
is that I think it's
important-- now,

565
00:26:25,000 --> 00:26:29,800
clearly the area of anonymity
and the role of technology

566
00:26:29,800 --> 00:26:31,360
here is ill understood.

567
00:26:31,360 --> 00:26:34,240
And we wouldn't be having
this panel here if we actually

568
00:26:34,240 --> 00:26:36,370
understood the problem exactly.

569
00:26:36,370 --> 00:26:41,380
And one of the hopes when
I got involved in all this,

570
00:26:41,380 --> 00:26:45,340
was trying to understand
the role between technology,

571
00:26:45,340 --> 00:26:47,290
and the role between
law enforcement,

572
00:26:47,290 --> 00:26:49,000
and the role between society.

573
00:26:49,000 --> 00:26:52,135
And if we hadn't done this
particular experiment,

574
00:26:52,135 --> 00:26:54,260
again, we wouldn't be
sitting here, and discussing,

575
00:26:54,260 --> 00:26:55,960
and having this
particular debate.

576
00:26:55,960 --> 00:26:59,530
And so to sum up, there's
three important reasons for me

577
00:26:59,530 --> 00:27:03,820
to be involved and
support the nym.alias.net.

578
00:27:03,820 --> 00:27:05,320
One is the freedom of speech.

579
00:27:05,320 --> 00:27:06,910
I'm a strong believer in it.

580
00:27:06,910 --> 00:27:09,190
Two, there are very
interesting research problems

581
00:27:09,190 --> 00:27:10,750
and we've got to
figure them out.

582
00:27:10,750 --> 00:27:14,590
Three, what's the role
of technology in society?

583
00:27:14,590 --> 00:27:20,470
And all of the three are
important in supporting

584
00:27:20,470 --> 00:27:21,890
this particular service.

585
00:27:21,890 --> 00:27:24,130
Now, Frans, you said
that it's actually not

586
00:27:24,130 --> 00:27:26,188
so hard to deal
with some problems

587
00:27:26,188 --> 00:27:28,480
that Phil brought up, some
crimes that Phil brought up,

588
00:27:28,480 --> 00:27:29,897
namely harassment,
that you've got

589
00:27:29,897 --> 00:27:31,840
a method for dealing with that.

590
00:27:31,840 --> 00:27:33,858
But you also said that
you've been the subject

591
00:27:33,858 --> 00:27:34,900
of a number of subpoenas.

592
00:27:34,900 --> 00:27:36,317
So it must be that
you can't quite

593
00:27:36,317 --> 00:27:40,840
deal with all the problems that
are at least allegedly arising

594
00:27:40,840 --> 00:27:41,710
out of your service.

595
00:27:41,710 --> 00:27:43,480
Tell us a little bit
about what happens

596
00:27:43,480 --> 00:27:44,480
when you get a subpoena.

597
00:27:44,480 --> 00:27:46,480
What sort of information
is asked for?

598
00:27:46,480 --> 00:27:50,102
What kind of information do
you give law enforcement?

599
00:27:50,102 --> 00:27:52,310
Let me make a couple of
statements along those lines.

600
00:27:52,310 --> 00:27:55,690
So first of all, the policy
statement that we actually

601
00:27:55,690 --> 00:27:59,200
have on nym.alias.net and
for any user on nym.alias.net

602
00:27:59,200 --> 00:28:02,020
says clearly that one
shouldn't use nym.alias.net

603
00:28:02,020 --> 00:28:06,350
for breaking any federal
or Massachusetts laws.

604
00:28:06,350 --> 00:28:10,480
And we have received
in the last couple

605
00:28:10,480 --> 00:28:12,040
of years, a number
of subpoenas where

606
00:28:12,040 --> 00:28:17,770
allegedly one of our users
did something incorrectly.

607
00:28:17,770 --> 00:28:23,010
And if we discuss
this with our lawyers,

608
00:28:23,010 --> 00:28:25,440
and if our lawyers
advise us to comply,

609
00:28:25,440 --> 00:28:27,270
we comply with the subpoena.

610
00:28:27,270 --> 00:28:30,660
And in most cases,
this means that there's

611
00:28:30,660 --> 00:28:35,640
a little bit of information
stored on nym.alias.net that

612
00:28:35,640 --> 00:28:39,030
basically says, if you think
about the chain of remailers,

613
00:28:39,030 --> 00:28:41,940
it basically says, what's
the last remailer where

614
00:28:41,940 --> 00:28:45,510
it came from, and where
it has to go to on its way

615
00:28:45,510 --> 00:28:47,070
back to the owner?

616
00:28:47,070 --> 00:28:52,440
And all this information
is carefully encrypted.

617
00:28:52,440 --> 00:28:54,780
The last hop has to be known.

618
00:28:54,780 --> 00:28:58,650
So we supply what we call the
reply block, which contains

619
00:28:58,650 --> 00:29:01,308
some information, a set
of instructions what

620
00:29:01,308 --> 00:29:03,600
the next hop should you do,
and the set of instructions

621
00:29:03,600 --> 00:29:05,820
are encrypted, so we don't
really know what it means,

622
00:29:05,820 --> 00:29:10,150
but we supply that particular
piece of information.

623
00:29:10,150 --> 00:29:11,675
So if a user--

624
00:29:11,675 --> 00:29:13,050
and this is the
important point--

625
00:29:13,050 --> 00:29:17,100
if a user uses our
service correctly,

626
00:29:17,100 --> 00:29:20,310
the information
that we're providing

627
00:29:20,310 --> 00:29:25,920
doesn't really have much value
to law enforcement, at least.

628
00:29:25,920 --> 00:29:27,690
Not clear what you
could learn from it.

629
00:29:27,690 --> 00:29:30,720
Of course, there are people
that use our system incorrectly

630
00:29:30,720 --> 00:29:33,840
and law enforcement
might actually

631
00:29:33,840 --> 00:29:35,700
have good luck with
this particular piece

632
00:29:35,700 --> 00:29:36,490
of information.

633
00:29:36,490 --> 00:29:38,590
So Phil, what do you
expect Frans to do?

634
00:29:38,590 --> 00:29:42,120
I mean, these guys can
obviously configure the system

635
00:29:42,120 --> 00:29:43,920
in a variety of ways.

636
00:29:43,920 --> 00:29:48,810
Frans has just told us that,
in fact, correctly used,

637
00:29:48,810 --> 00:29:50,400
his system doesn't
really provide

638
00:29:50,400 --> 00:29:52,110
much useful information.

639
00:29:52,110 --> 00:29:54,030
What are these guys to do?

640
00:29:54,030 --> 00:29:57,300
Well, I don't have any
real silver bullets, Danny.

641
00:29:57,300 --> 00:30:00,330
I hate to tell you
that, but it's true.

642
00:30:00,330 --> 00:30:06,910
It seems to me that people
are too ready, generally,

643
00:30:06,910 --> 00:30:12,730
to resort to law to solve
all of the problems presented

644
00:30:12,730 --> 00:30:14,180
by technology.

645
00:30:14,180 --> 00:30:16,120
Technology can solve
a lot of the problems,

646
00:30:16,120 --> 00:30:19,240
policies and practices can
solve a lot of the problems.

647
00:30:19,240 --> 00:30:25,480
And if I may refer to the
nym.alias.net policy FAQ,

648
00:30:25,480 --> 00:30:26,680
it has a great line in it.

649
00:30:26,680 --> 00:30:29,110
"Out experience with
nym.alias.net shows

650
00:30:29,110 --> 00:30:31,540
that controlling
abuse is as important

651
00:30:31,540 --> 00:30:34,640
as protecting the identities
of anonymous users",

652
00:30:34,640 --> 00:30:37,120
excuse me, users.

653
00:30:37,120 --> 00:30:40,570
I think the first and primary
thing that has to happen

654
00:30:40,570 --> 00:30:44,590
is that people who are
responsible for designing

655
00:30:44,590 --> 00:30:47,410
and implementing
technology and standards

656
00:30:47,410 --> 00:30:52,000
need to be sensitive to how
technologies can be abused,

657
00:30:52,000 --> 00:30:56,770
and need to consider
the extent to which that

658
00:30:56,770 --> 00:31:02,820
can happen in designing and
implementing the systems.

659
00:31:02,820 --> 00:31:06,270
I agree completely, by
the way, with Nadine,

660
00:31:06,270 --> 00:31:09,540
that we should not reason
from worst-case scenarios.

661
00:31:09,540 --> 00:31:12,030
Either the worst-case
scenario, probably,

662
00:31:12,030 --> 00:31:17,190
of the kidnapper who you need
to trace immediately or people

663
00:31:17,190 --> 00:31:18,450
will die.

664
00:31:18,450 --> 00:31:21,180
But you also don't
need to reason

665
00:31:21,180 --> 00:31:23,760
from the worst-case
scenario on the other side

666
00:31:23,760 --> 00:31:27,390
of the human rights
advocate in a country who,

667
00:31:27,390 --> 00:31:30,930
if his identity is exposed,
will be immediately executed.

668
00:31:30,930 --> 00:31:32,820
These these are
both harms that you

669
00:31:32,820 --> 00:31:35,490
have to take into
consideration in developing

670
00:31:35,490 --> 00:31:39,030
a balanced policy that addresses
all of the relevant concerns.

671
00:31:39,030 --> 00:31:42,900
I think Phil's point about
a lot of these problems,

672
00:31:42,900 --> 00:31:46,200
we should see technology
as providing solutions

673
00:31:46,200 --> 00:31:50,940
and not jump to law, I'd
like to agree with that point

674
00:31:50,940 --> 00:31:53,610
but come at it from a
different perspective, which

675
00:31:53,610 --> 00:31:59,790
is, I'm very concerned about
the very enormous threats

676
00:31:59,790 --> 00:32:04,620
to individual privacy that have
occurred in cyberspace because

677
00:32:04,620 --> 00:32:09,480
of technological power
that has vastly increased

678
00:32:09,480 --> 00:32:12,540
the ability of everybody,
from government

679
00:32:12,540 --> 00:32:17,940
to the private sector, to
engage in snooping, and putting

680
00:32:17,940 --> 00:32:21,960
together, and trading in
what many of us like to think

681
00:32:21,960 --> 00:32:26,100
is personal information that
pertains only to us, including

682
00:32:26,100 --> 00:32:28,590
information of a
very personal nature,

683
00:32:28,590 --> 00:32:32,760
such as financial information
and medical information.

684
00:32:32,760 --> 00:32:35,350
Unfortunately, from
my point of view,

685
00:32:35,350 --> 00:32:37,620
the laws in the
United States are not

686
00:32:37,620 --> 00:32:41,130
very protective of individual
privacy with respect

687
00:32:41,130 --> 00:32:42,960
to that kind of information.

688
00:32:42,960 --> 00:32:46,980
And therefore, I welcome
the kind of technology

689
00:32:46,980 --> 00:32:50,910
that LCS has put together
through the anonymous remailer,

690
00:32:50,910 --> 00:32:54,760
together with other such
privacy protecting technologies,

691
00:32:54,760 --> 00:32:59,760
including encryption, as the
only way that individuals

692
00:32:59,760 --> 00:33:03,840
can try to enjoy, in
cyberspace, the kind of privacy

693
00:33:03,840 --> 00:33:08,820
that we've traditionally
enjoyed in other kinds of media.

694
00:33:08,820 --> 00:33:11,140
So Nadine, you're
a law professor,

695
00:33:11,140 --> 00:33:13,500
you're a defender
of the Constitution,

696
00:33:13,500 --> 00:33:16,920
and you appear to be ready to
throw it all out the window

697
00:33:16,920 --> 00:33:19,170
and rely on what
these guys sitting up

698
00:33:19,170 --> 00:33:21,570
at the table and our
colleagues at in the audience

699
00:33:21,570 --> 00:33:22,195
have developed.

700
00:33:22,195 --> 00:33:24,870
I'd rather rely on them than the
current United States Supreme

701
00:33:24,870 --> 00:33:26,738
Court, I'm sorry to say.

702
00:33:26,738 --> 00:33:28,112
[LAUGHS]

703
00:33:28,112 --> 00:33:30,860
[APPLAUSE]

704
00:33:30,860 --> 00:33:35,910
But in all seriousness,
I mean, we do have,

705
00:33:35,910 --> 00:33:38,520
as you well know, a
Fourth Amendment to the US

706
00:33:38,520 --> 00:33:41,338
Constitution that protects
against unreasonable search

707
00:33:41,338 --> 00:33:41,880
and seizures.

708
00:33:41,880 --> 00:33:45,240
We have an extensive
wiretapping law

709
00:33:45,240 --> 00:33:48,630
which prescribes very
difficult hurdles

710
00:33:48,630 --> 00:33:51,870
that Phil and his colleagues
have to jump through in order

711
00:33:51,870 --> 00:33:55,260
to gain access to
private information,

712
00:33:55,260 --> 00:33:59,790
whether on the internet or other
telecommunications systems.

713
00:33:59,790 --> 00:34:01,440
Isn't that enough for you?

714
00:34:01,440 --> 00:34:04,830
Well, it would be if the
Fourth Amendment were

715
00:34:04,830 --> 00:34:07,350
enforced in accordance
with its plain language

716
00:34:07,350 --> 00:34:08,460
and original intent.

717
00:34:08,460 --> 00:34:09,991
But that's, in
fact, not the case.

718
00:34:09,991 --> 00:34:11,699
We're sounding very
conservative tonight.

719
00:34:11,699 --> 00:34:14,280
Really conservative.

720
00:34:14,280 --> 00:34:16,710
About almost 20 years
ago, a law professor

721
00:34:16,710 --> 00:34:19,770
wrote an article called, The
Incredible Shrinking Fourth

722
00:34:19,770 --> 00:34:20,580
Amendment.

723
00:34:20,580 --> 00:34:23,760
And unfortunately, it's
shrunk a lot more since then.

724
00:34:23,760 --> 00:34:27,330
Most recently, just last week,
the United States Supreme Court

725
00:34:27,330 --> 00:34:28,620
cut back very far.

726
00:34:28,620 --> 00:34:30,630
And I think the most
pernicious thing that's

727
00:34:30,630 --> 00:34:33,420
happened to constitutional
privacy doctrine

728
00:34:33,420 --> 00:34:37,350
is that rather than setting
an absolute guarantee, which

729
00:34:37,350 --> 00:34:40,770
I believe the Fourth Amendment
was clearly intended to do,

730
00:34:40,770 --> 00:34:43,110
it's become extremely relative.

731
00:34:43,110 --> 00:34:46,139
And this started happening
about a generation ago

732
00:34:46,139 --> 00:34:49,110
when the Supreme Court
said, you will only

733
00:34:49,110 --> 00:34:53,760
be protected in what you claim
to be a privacy right if it

734
00:34:53,760 --> 00:34:58,560
is something that
society recognizes

735
00:34:58,560 --> 00:35:02,860
where you have a legitimate
expectation of privacy.

736
00:35:02,860 --> 00:35:07,020
So if you know that your
government, together

737
00:35:07,020 --> 00:35:10,170
with private actors, are
putting video surveillance

738
00:35:10,170 --> 00:35:13,530
cameras in public
places, if you know

739
00:35:13,530 --> 00:35:16,440
that websites are
using cookies, if you

740
00:35:16,440 --> 00:35:19,800
know that your employer is
engaging in wiretapping,

741
00:35:19,800 --> 00:35:24,720
you no longer have a legitimate
expectation of privacy,

742
00:35:24,720 --> 00:35:26,250
says the Supreme Court.

743
00:35:26,250 --> 00:35:28,880
Sorry, the Fourth
Amendment doesn't apply.

744
00:35:28,880 --> 00:35:34,050
So this vicious cycle has set up
that the more we have invasions

745
00:35:34,050 --> 00:35:38,220
of privacy, the less we have
constitutional protections

746
00:35:38,220 --> 00:35:40,290
against it.

747
00:35:40,290 --> 00:35:42,240
I want to get Phil's
response, because we

748
00:35:42,240 --> 00:35:46,680
seem to have backed ourselves
into a vicious circle where

749
00:35:46,680 --> 00:35:50,250
the lawyers are concerned about
what the technologists are

750
00:35:50,250 --> 00:35:53,700
doing, and the users
of the technology

751
00:35:53,700 --> 00:35:55,620
don't trust the legal
system to protect them,

752
00:35:55,620 --> 00:35:58,510
so they're heading
back to the technology.

753
00:35:58,510 --> 00:36:02,000
Well, if I could respond, then.

754
00:36:02,000 --> 00:36:03,830
Yes.

755
00:36:03,830 --> 00:36:05,840
You can get us out of
this vicious circle.

756
00:36:05,840 --> 00:36:08,810
I don't know if I can do that.

757
00:36:08,810 --> 00:36:13,940
I, as both a human being and
a member of the United States

758
00:36:13,940 --> 00:36:15,355
government, find--

759
00:36:15,355 --> 00:36:16,355
Two separate categories?

760
00:36:16,355 --> 00:36:18,985
[LAUGHTER]

761
00:36:18,985 --> 00:36:21,800


762
00:36:21,800 --> 00:36:23,315
They are some sort
of Venn diagram.

763
00:36:23,315 --> 00:36:25,880


764
00:36:25,880 --> 00:36:27,612
Whether one is
completely captured

765
00:36:27,612 --> 00:36:29,820
within the other or not,
I'll leave for you to judge.

766
00:36:29,820 --> 00:36:32,480
I think there are no non humans
employed by the United States

767
00:36:32,480 --> 00:36:35,210
government, it's fair to say.

768
00:36:35,210 --> 00:36:37,230
Privacy is very important to me.

769
00:36:37,230 --> 00:36:39,230
But I am not--

770
00:36:39,230 --> 00:36:41,330
Drug sniffing dogs.

771
00:36:41,330 --> 00:36:44,660
But privacy is not
an absolute right.

772
00:36:44,660 --> 00:36:47,030
It's one of a lot
of rights which have

773
00:36:47,030 --> 00:36:48,590
to be taken into consideration.

774
00:36:48,590 --> 00:36:51,110
If somebody broke
into your house

775
00:36:51,110 --> 00:36:53,540
and stole everything
you owned, would you

776
00:36:53,540 --> 00:36:56,090
defend the right of that
person to remain private?

777
00:36:56,090 --> 00:36:57,710
Probably not.

778
00:36:57,710 --> 00:37:01,898
We have, from the beginning
of this country, recognized--

779
00:37:01,898 --> 00:37:03,440
and the Fourth
Amendment recognizes--

780
00:37:03,440 --> 00:37:06,140
that there are times
when we can interrupt

781
00:37:06,140 --> 00:37:08,820
people's reasonable
expectations of privacy,

782
00:37:08,820 --> 00:37:11,450
particularly by getting a
search warrant that allows law

783
00:37:11,450 --> 00:37:14,570
enforcement to go in and
discover the things it needs

784
00:37:14,570 --> 00:37:17,510
to do to enforce the law.

785
00:37:17,510 --> 00:37:21,730
This is why anonymity is
also not an absolute right.

786
00:37:21,730 --> 00:37:23,570
Frans here, was
wearing a mask before.

787
00:37:23,570 --> 00:37:25,310
This may come as a
shock to you, he's

788
00:37:25,310 --> 00:37:27,320
not allowed to do that
in some courthouses.

789
00:37:27,320 --> 00:37:28,970
And the Seventh
Circuit, at least,

790
00:37:28,970 --> 00:37:32,210
has said that laws banning him
wearing a mask in a courthouse

791
00:37:32,210 --> 00:37:34,430
is OK.

792
00:37:34,430 --> 00:37:39,260
We don't allow people to drive
anonymously in the real world.

793
00:37:39,260 --> 00:37:43,700
We have to recognize that there
are costs to allowing people,

794
00:37:43,700 --> 00:37:46,700
allowing the government to
breach anonymity and find out

795
00:37:46,700 --> 00:37:48,290
who is doing what.

796
00:37:48,290 --> 00:37:50,570
There are benefits
to doing that.

797
00:37:50,570 --> 00:37:53,120
If you give government
the ability to do that,

798
00:37:53,120 --> 00:37:55,070
there will be abuses.

799
00:37:55,070 --> 00:37:58,070
If you don't give government
the ability to do that,

800
00:37:58,070 --> 00:38:00,320
then there will be
no law enforcement.

801
00:38:00,320 --> 00:38:03,140
That's why you have to have
a balanced solution that

802
00:38:03,140 --> 00:38:05,600
takes into account what
technology can provide

803
00:38:05,600 --> 00:38:08,540
and what policies
can do and address

804
00:38:08,540 --> 00:38:11,690
the balance with an
proprioceptive of standards.

805
00:38:11,690 --> 00:38:16,250
The two questions really
are, should law enforcement

806
00:38:16,250 --> 00:38:19,280
be able to trace
communications on the internet

807
00:38:19,280 --> 00:38:21,450
in order to enforce the law?

808
00:38:21,450 --> 00:38:23,270
And if they should--

809
00:38:23,270 --> 00:38:24,770
I think the answer
is they should,

810
00:38:24,770 --> 00:38:27,228
that we shouldn't have no law
enforcement on the internet--

811
00:38:27,228 --> 00:38:30,500
what ought the standards
be for law enforcement

812
00:38:30,500 --> 00:38:33,080
to get access to the
information in order

813
00:38:33,080 --> 00:38:34,630
to trace communication.

814
00:38:34,630 --> 00:38:35,030
Can I just interject
something about--

815
00:38:35,030 --> 00:38:36,572
Please, and then
we'll come to David.

816
00:38:36,572 --> 00:38:38,870
Because the United
States Supreme Court did,

817
00:38:38,870 --> 00:38:40,730
two years ago, issue--

818
00:38:40,730 --> 00:38:42,080
four years ago, I'm sorry--

819
00:38:42,080 --> 00:38:43,860
issue a decision.

820
00:38:43,860 --> 00:38:48,500
7 to 2 was the vote, so very
lopsided in favor of the right

821
00:38:48,500 --> 00:38:51,800
to engage in
anonymous expression.

822
00:38:51,800 --> 00:38:54,255
The Seventh Circuit
decision was wrong,

823
00:38:54,255 --> 00:38:56,130
but that's not the
highest court in the land.

824
00:38:56,130 --> 00:38:57,930
We won't get there.

825
00:38:57,930 --> 00:39:01,250
And in that case, the
United States Supreme Court

826
00:39:01,250 --> 00:39:04,820
said, yes, it's not
an absolute right,

827
00:39:04,820 --> 00:39:08,900
but it is a very strongly
protected right, which

828
00:39:08,900 --> 00:39:13,880
means it may be limited if, but
only if, government can meet

829
00:39:13,880 --> 00:39:17,000
a very high standard of proof.

830
00:39:17,000 --> 00:39:19,340
Namely, it has to
show an interest

831
00:39:19,340 --> 00:39:21,740
of compelling importance.

832
00:39:21,740 --> 00:39:25,880
Thinking that Frans has a cute
face that we'd like to look at

833
00:39:25,880 --> 00:39:27,680
is not of compelling importance.

834
00:39:27,680 --> 00:39:31,250
Obviously, dealing with
a serious crime would be.

835
00:39:31,250 --> 00:39:32,160
Sorry.

836
00:39:32,160 --> 00:39:34,460
[LAUGHS]

837
00:39:34,460 --> 00:39:38,120
But even more importantly
is the second requirement.

838
00:39:38,120 --> 00:39:42,920
And that is that there is no
less restrictive alternative

839
00:39:42,920 --> 00:39:46,070
means that government
can use to promote

840
00:39:46,070 --> 00:39:47,480
that compelling purpose.

841
00:39:47,480 --> 00:39:49,940
So this case that
I'm talking about

842
00:39:49,940 --> 00:39:52,490
was an ACLU case, by the way.

843
00:39:52,490 --> 00:39:57,560
We defended the right of a woman
to hand out anonymous handbills

844
00:39:57,560 --> 00:40:00,950
or distribute anonymous
campaign literature with respect

845
00:40:00,950 --> 00:40:04,610
to a campaign issue,
a political issue.

846
00:40:04,610 --> 00:40:07,430
And the government had,
the Supreme Court agreed,

847
00:40:07,430 --> 00:40:09,530
a compellingly
important interest

848
00:40:09,530 --> 00:40:13,190
in being sure that there
wasn't false information

849
00:40:13,190 --> 00:40:15,440
injected into the
political campaign,

850
00:40:15,440 --> 00:40:19,130
but it said laws could be
written that more narrowly

851
00:40:19,130 --> 00:40:22,280
would deal with that problem,
laws specifically targeting

852
00:40:22,280 --> 00:40:23,630
fraud or deception.

853
00:40:23,630 --> 00:40:25,760
So I would think
at the very least--

854
00:40:25,760 --> 00:40:28,070
and maybe this is
not fair-- but I

855
00:40:28,070 --> 00:40:29,780
would think at the
very least, Phil,

856
00:40:29,780 --> 00:40:32,360
you would agree that the
broad question that's

857
00:40:32,360 --> 00:40:36,530
posed to this panel, should
the anonymous remailer be shut

858
00:40:36,530 --> 00:40:39,620
down, would have to be
answered in the negative

859
00:40:39,620 --> 00:40:44,240
because your concern has to
do with the kind of abuses

860
00:40:44,240 --> 00:40:46,010
that, as the Supreme
Court has said,

861
00:40:46,010 --> 00:40:49,310
you don't deal with by
completely prohibiting

862
00:40:49,310 --> 00:40:50,060
anonymity.

863
00:40:50,060 --> 00:40:53,510
You deal with it in a more
selected, targeted, focused

864
00:40:53,510 --> 00:40:54,740
fashion.

865
00:40:54,740 --> 00:40:57,590
Well, it seems that one
of the dilemmas we have

866
00:40:57,590 --> 00:41:01,110
is that the technology--

867
00:41:01,110 --> 00:41:03,030
I think we heard from David--

868
00:41:03,030 --> 00:41:06,000
is not all that good
at being selective.

869
00:41:06,000 --> 00:41:08,760
In other words,
these guys have not

870
00:41:08,760 --> 00:41:11,970
built a political speech
anonymous remailer

871
00:41:11,970 --> 00:41:15,060
that rejects any speech
which is not political.

872
00:41:15,060 --> 00:41:20,060
They've not built the remailer
for the late Mrs. McIntyre.

873
00:41:20,060 --> 00:41:24,570
Not-- yeah, McIntyre, who
was the plaintiff in the case

874
00:41:24,570 --> 00:41:28,890
that you mentioned, they
built a remailer for everyone.

875
00:41:28,890 --> 00:41:31,080
For Mrs. McIntyre,
for the person

876
00:41:31,080 --> 00:41:33,990
who would harass Mrs.
McIntyre, for the person who

877
00:41:33,990 --> 00:41:36,270
would blow up the city
that Mrs. McIntyre

878
00:41:36,270 --> 00:41:38,350
lives in, wherever that is.

879
00:41:38,350 --> 00:41:40,890
So David, I guess
the question to you,

880
00:41:40,890 --> 00:41:44,040
since you've built
this stuff, is, are you

881
00:41:44,040 --> 00:41:45,960
going to help Phil out?

882
00:41:45,960 --> 00:41:50,940
Phil has said that he's going to
be extra, extra, extra careful.

883
00:41:50,940 --> 00:41:53,440
And Nadine is probably
going to keep Phil honest.

884
00:41:53,440 --> 00:41:57,480
She's going to go into court
time, and time, and time again.

885
00:41:57,480 --> 00:42:01,830
When Phil shows up at your
door with a search warrant,

886
00:42:01,830 --> 00:42:04,860
with a wiretap order for
access to information,

887
00:42:04,860 --> 00:42:06,930
Nadine is a great
lawyer, is going

888
00:42:06,930 --> 00:42:08,700
to figure out how
to find out what

889
00:42:08,700 --> 00:42:10,080
it is that Phil is asking for.

890
00:42:10,080 --> 00:42:12,720
And if he's asking
for too much, the ACLU

891
00:42:12,720 --> 00:42:16,080
or some other organization,
the defense attorney

892
00:42:16,080 --> 00:42:17,310
is going to stop Phil.

893
00:42:17,310 --> 00:42:19,380
So are you going
to build something

894
00:42:19,380 --> 00:42:22,970
that gives Phil a little bit?

895
00:42:22,970 --> 00:42:28,190
Well, I think in some sense I'm
interested in helping Phil out,

896
00:42:28,190 --> 00:42:31,880
but I'm coming at it from a very
different angle in the sense

897
00:42:31,880 --> 00:42:34,513
that I'm very interested
in making sure

898
00:42:34,513 --> 00:42:35,930
that anonymous
remailers, in fact,

899
00:42:35,930 --> 00:42:39,958
can't be used to do horrible
things, and commit crimes,

900
00:42:39,958 --> 00:42:40,500
and so forth.

901
00:42:40,500 --> 00:42:43,580
And in particular, probably
one of the greatest threats

902
00:42:43,580 --> 00:42:47,180
of an anonymous remailer is
in terms of computer security,

903
00:42:47,180 --> 00:42:49,970
potentially attacking
computer systems.

904
00:42:49,970 --> 00:42:56,270
And so certainly, to
the extent that we

905
00:42:56,270 --> 00:42:59,810
can do things to prevent
the kinds of crimes

906
00:42:59,810 --> 00:43:02,580
that you might be worried
about, I'm all in favor of that.

907
00:43:02,580 --> 00:43:05,330
And in particular, if you're
worried about the White House

908
00:43:05,330 --> 00:43:07,640
receiving a lot of
bomb threats, I'm

909
00:43:07,640 --> 00:43:10,140
perfectly happy if the White
House comes to me and says,

910
00:43:10,140 --> 00:43:13,307
well, you're costing the Secret
Service hundreds of thousands

911
00:43:13,307 --> 00:43:15,140
of dollars to search
all these bomb threats,

912
00:43:15,140 --> 00:43:16,520
and you know this
president isn't

913
00:43:16,520 --> 00:43:19,190
interested in getting anonymous
political advice anyway,

914
00:43:19,190 --> 00:43:20,750
just stop sending
us mail, right?

915
00:43:20,750 --> 00:43:21,680
OK, no problem.

916
00:43:21,680 --> 00:43:24,138
You're not going to receive
another piece of anonymous mail

917
00:43:24,138 --> 00:43:25,850
from our server if
you don't want to.

918
00:43:25,850 --> 00:43:28,940
But again, that's up to you.

919
00:43:28,940 --> 00:43:33,230
I mean, one thing I'd like to
add, going back a little bit,

920
00:43:33,230 --> 00:43:36,710
is that I shared
Nadine's concern

921
00:43:36,710 --> 00:43:39,440
that the Fourth Amendment
may be shrinking.

922
00:43:39,440 --> 00:43:42,290
But in the particular case
of this anonymous remailer,

923
00:43:42,290 --> 00:43:45,890
the US government is, in
fact, not the only enemy.

924
00:43:45,890 --> 00:43:49,003
And one thing that has
sort of become apparent

925
00:43:49,003 --> 00:43:51,170
is that if you have anonymous
systems and they fail,

926
00:43:51,170 --> 00:43:52,580
people can get hurt.

927
00:43:52,580 --> 00:43:58,932
And so one thing that I'm
interested in this research,

928
00:43:58,932 --> 00:44:00,140
and I want to carry this out.

929
00:44:00,140 --> 00:44:01,665
And in order to
carry this out, I

930
00:44:01,665 --> 00:44:04,040
need to really deploy the
system in the real world, which

931
00:44:04,040 --> 00:44:06,610
means that there are people who
are relying on this system--

932
00:44:06,610 --> 00:44:07,610
Who's going to get hurt?

933
00:44:07,610 --> 00:44:08,150
--for their safety.

934
00:44:08,150 --> 00:44:09,150
Who's going to get hurt?

935
00:44:09,150 --> 00:44:12,413
Well, for example,
people in countries

936
00:44:12,413 --> 00:44:13,580
with oppressive governments.

937
00:44:13,580 --> 00:44:16,550
Or people who have escaped
cults, for example,

938
00:44:16,550 --> 00:44:18,922
and are trying to help
other people escape cults.

939
00:44:18,922 --> 00:44:20,630
There's some cults
where in order to join

940
00:44:20,630 --> 00:44:22,547
you have to give a lot
of personal information

941
00:44:22,547 --> 00:44:23,360
about yourself.

942
00:44:23,360 --> 00:44:24,980
And you can be
seriously harassed

943
00:44:24,980 --> 00:44:29,330
afterwards if you start trying
to work against the cult.

944
00:44:29,330 --> 00:44:32,780
And in particular,
I have an office

945
00:44:32,780 --> 00:44:34,730
with a standard MIT
lock on the door,

946
00:44:34,730 --> 00:44:37,050
and I'm sure it's
not very secure.

947
00:44:37,050 --> 00:44:39,032
And if I had some
machine in my office

948
00:44:39,032 --> 00:44:41,240
that actually had information
that was of great value

949
00:44:41,240 --> 00:44:43,228
to someone, say, other
than the US government,

950
00:44:43,228 --> 00:44:45,020
say some organization
that didn't, in fact,

951
00:44:45,020 --> 00:44:46,520
care about respecting
the laws, then

952
00:44:46,520 --> 00:44:48,937
I'd be worried that someone
could just walk into my office

953
00:44:48,937 --> 00:44:50,330
and walk out with the machine.

954
00:44:50,330 --> 00:44:54,788
And so, if in fact, I have to
worry a lot about the security

955
00:44:54,788 --> 00:44:57,080
of the system or worry about
potentially getting people

956
00:44:57,080 --> 00:44:59,510
hurt, then that's going
to be a serious impediment

957
00:44:59,510 --> 00:45:01,560
to my research.

958
00:45:01,560 --> 00:45:07,680
Now, Phil, you've urged
us to rely on the,

959
00:45:07,680 --> 00:45:09,540
I think what we
would all agree is

960
00:45:09,540 --> 00:45:13,290
the very delicate
and careful balance

961
00:45:13,290 --> 00:45:16,590
that the Fourth
Amendment strikes here.

962
00:45:16,590 --> 00:45:20,520
Michael Dertouzos, ever
internationally-minded has

963
00:45:20,520 --> 00:45:21,540
pointed out that--

964
00:45:21,540 --> 00:45:22,620
what did he say?

965
00:45:22,620 --> 00:45:27,060
Only 4% of the world is
in the United States.

966
00:45:27,060 --> 00:45:29,430
These guys have
built the technology

967
00:45:29,430 --> 00:45:33,240
to help not just human rights
workers in other countries,

968
00:45:33,240 --> 00:45:36,120
but people who live in
countries which do not

969
00:45:36,120 --> 00:45:40,230
have the benefit of the
Fourth Amendment protections

970
00:45:40,230 --> 00:45:42,360
that we have as US citizens.

971
00:45:42,360 --> 00:45:46,080
Now, all those people
are not necessarily

972
00:45:46,080 --> 00:45:50,130
your responsibility, and I don't
want to put it on you alone.

973
00:45:50,130 --> 00:45:54,450
But if you are urging
the design of systems

974
00:45:54,450 --> 00:45:58,710
to strike a more careful
balance between US law,

975
00:45:58,710 --> 00:46:01,020
and the rights that we
have as US citizens,

976
00:46:01,020 --> 00:46:04,200
and the needs of law enforcement
in a Democratic country

977
00:46:04,200 --> 00:46:08,670
like ours, what does that
do in a global environment

978
00:46:08,670 --> 00:46:12,850
to people who don't have those
protections available to them?

979
00:46:12,850 --> 00:46:15,670
Well, I'll steal an
argument from Nadine,

980
00:46:15,670 --> 00:46:17,380
I think that argument
may prove too much.

981
00:46:17,380 --> 00:46:20,130


982
00:46:20,130 --> 00:46:24,060
Police officers in other
countries may use firearms

983
00:46:24,060 --> 00:46:26,820
to kill people that
they shouldn't.

984
00:46:26,820 --> 00:46:29,460
That doesn't mean we don't allow
police officers in the United

985
00:46:29,460 --> 00:46:31,380
States to be armed.

986
00:46:31,380 --> 00:46:34,440
The fact that a technology can
be abused outside the United

987
00:46:34,440 --> 00:46:37,470
States is a fact to be
taken into consideration

988
00:46:37,470 --> 00:46:40,020
in deciding what the
appropriate balance is.

989
00:46:40,020 --> 00:46:41,620
But it's just an
additional fact.

990
00:46:41,620 --> 00:46:44,250
You still have to come up
with a balanced policy.

991
00:46:44,250 --> 00:46:46,140
And in particular,
where those systems are

992
00:46:46,140 --> 00:46:49,310
located in the United States.

993
00:46:49,310 --> 00:46:51,320
The anonymous
remailer is not going

994
00:46:51,320 --> 00:46:54,770
to have to respond to requests
from foreign governments.

995
00:46:54,770 --> 00:46:56,270
The requests are
going to have to be

996
00:46:56,270 --> 00:46:58,430
root through the US government.

997
00:46:58,430 --> 00:47:01,460
And the US government is
going to make its own decision

998
00:47:01,460 --> 00:47:03,440
on whether it's
appropriate to bring

999
00:47:03,440 --> 00:47:06,770
legal process in
the United States,

1000
00:47:06,770 --> 00:47:08,870
if it's a foreign law
enforcement request.

1001
00:47:08,870 --> 00:47:11,960
A more serious threat
to anonymous remailers

1002
00:47:11,960 --> 00:47:14,660
may be civil lawsuits
in the United States

1003
00:47:14,660 --> 00:47:18,090
about people who want
access to information.

1004
00:47:18,090 --> 00:47:18,590
[INAUDIBLE]

1005
00:47:18,590 --> 00:47:19,820
Can you give us an example?

1006
00:47:19,820 --> 00:47:23,060
There's a fairly
recent published case

1007
00:47:23,060 --> 00:47:29,060
involving an internet
service provider where a--

1008
00:47:29,060 --> 00:47:31,400
[INAUDIBLE] if I remember--

1009
00:47:31,400 --> 00:47:33,500
a couple was getting divorced.

1010
00:47:33,500 --> 00:47:39,080
The new spouse of
the divorced man

1011
00:47:39,080 --> 00:47:42,290
published some
derogatory information

1012
00:47:42,290 --> 00:47:46,190
using a screen name
on the internet

1013
00:47:46,190 --> 00:47:48,200
about the prior spouse.

1014
00:47:48,200 --> 00:47:52,910
And the spouse who
was defamed was

1015
00:47:52,910 --> 00:47:55,880
able to get the information
about who did that

1016
00:47:55,880 --> 00:47:58,980
from the service provider,
because there were no anonymity

1017
00:47:58,980 --> 00:47:59,480
protections.

1018
00:47:59,480 --> 00:48:01,710
It was just done
under a screen name.

1019
00:48:01,710 --> 00:48:04,802
Now, you can argue
that either way,

1020
00:48:04,802 --> 00:48:07,010
whether you think that's a
good thing or a bad thing.

1021
00:48:07,010 --> 00:48:09,190
The ability was
there in that case.

1022
00:48:09,190 --> 00:48:10,960
And you can also argue--

1023
00:48:10,960 --> 00:48:12,940
I'm sure Nadine would
say, well, there

1024
00:48:12,940 --> 00:48:15,190
weren't sufficient
protections for that person's

1025
00:48:15,190 --> 00:48:16,810
anonymous identification.

1026
00:48:16,810 --> 00:48:18,310
She might be right.

1027
00:48:18,310 --> 00:48:20,740
But I think the way you
address that type of problem

1028
00:48:20,740 --> 00:48:25,180
is by setting legal standards
for access to the information.

1029
00:48:25,180 --> 00:48:27,340
And I'm sure Nadine will
tell you that we already

1030
00:48:27,340 --> 00:48:30,442
have, with regard to
government access, incredibly

1031
00:48:30,442 --> 00:48:32,775
complicated standards under
the electronic communication

1032
00:48:32,775 --> 00:48:33,040
[INAUDIBLE].

1033
00:48:33,040 --> 00:48:33,707
Tell us, Nadine.

1034
00:48:33,707 --> 00:48:35,170
Well, I agree with that.

1035
00:48:35,170 --> 00:48:38,500
And they're not nearly as
protective as they should be.

1036
00:48:38,500 --> 00:48:41,290
But I want to come
back to Michael's point

1037
00:48:41,290 --> 00:48:44,300
as conveyed by Danny from a
somewhat different perspective,

1038
00:48:44,300 --> 00:48:47,080
which is, I want to be very
clear that I'm not here

1039
00:48:47,080 --> 00:48:51,520
advocating American imperialism
in terms of exporting

1040
00:48:51,520 --> 00:48:53,710
the Fourth Amendment
as originally written

1041
00:48:53,710 --> 00:48:54,460
around the world.

1042
00:48:54,460 --> 00:48:56,650
I certainly wouldn't
want to export it

1043
00:48:56,650 --> 00:48:59,600
in its current
watered down fashion.

1044
00:48:59,600 --> 00:49:04,990
But what we are talking about
are basic concepts of privacy

1045
00:49:04,990 --> 00:49:08,530
that are respected and
reflected at least as

1046
00:49:08,530 --> 00:49:11,740
strongly in international
human rights law,

1047
00:49:11,740 --> 00:49:14,680
and in regional human
rights agreements,

1048
00:49:14,680 --> 00:49:19,930
and in the constitutions of
every country around the world.

1049
00:49:19,930 --> 00:49:22,990
Now, as is often the
case in our own country,

1050
00:49:22,990 --> 00:49:26,410
sometimes those guarantees
are honored in the breach.

1051
00:49:26,410 --> 00:49:31,810
But it is a standard that
has been at least aspired to

1052
00:49:31,810 --> 00:49:34,900
by governments, and regional,
and international bodies

1053
00:49:34,900 --> 00:49:39,280
around the world, and is
advocated by cyber libertarians

1054
00:49:39,280 --> 00:49:40,420
around the world.

1055
00:49:40,420 --> 00:49:42,370
When I work on
these issues, it's

1056
00:49:42,370 --> 00:49:46,570
not only in coalition with other
organizations in this country,

1057
00:49:46,570 --> 00:49:49,810
but also organizations around
the world, many of which

1058
00:49:49,810 --> 00:49:53,920
are working together
under euphonious named

1059
00:49:53,920 --> 00:49:57,940
GILC, which stands for the--
the first time I heard it

1060
00:49:57,940 --> 00:49:59,530
I thought, geek.

1061
00:49:59,530 --> 00:50:00,220
Not quite.

1062
00:50:00,220 --> 00:50:03,960
It's the Global Internet
Liberty Campaign.

1063
00:50:03,960 --> 00:50:04,470
Frans?

1064
00:50:04,470 --> 00:50:06,095
Yeah, I just wanted
wanted to follow up

1065
00:50:06,095 --> 00:50:08,815
on this point of international
and globalization.

1066
00:50:08,815 --> 00:50:10,440
And I think this is
one of the problems

1067
00:50:10,440 --> 00:50:11,982
that, actually, Phil
is going to face

1068
00:50:11,982 --> 00:50:14,700
is, this particular service
runs in the United States,

1069
00:50:14,700 --> 00:50:17,490
but there's no reason
why there couldn't

1070
00:50:17,490 --> 00:50:19,590
be a number of them
outside of United States,

1071
00:50:19,590 --> 00:50:22,530
where the United States
law doesn't apply at all.

1072
00:50:22,530 --> 00:50:28,710
And in fact, work even harder
for you to try to, I guess,

1073
00:50:28,710 --> 00:50:31,470
create any particular
piece of communication.

1074
00:50:31,470 --> 00:50:33,150
So one of the
things that we have

1075
00:50:33,150 --> 00:50:35,160
to think about I think
when we move forward,

1076
00:50:35,160 --> 00:50:37,920
and as the change the
internet is bringing

1077
00:50:37,920 --> 00:50:42,840
is that we get this sort
of universal globalization

1078
00:50:42,840 --> 00:50:46,710
and everybody can run
these kinds of services.

1079
00:50:46,710 --> 00:50:49,740
And that changes, I think,
some of the perspectives.

1080
00:50:49,740 --> 00:50:52,650
Because even if we enact
some law here that says,

1081
00:50:52,650 --> 00:50:55,530
we decide today, or somebody
decides that nym.alias.net

1082
00:50:55,530 --> 00:50:58,030
shouldn't be running,
somebody can go off,

1083
00:50:58,030 --> 00:51:00,030
and actually, our software's
publicly available,

1084
00:51:00,030 --> 00:51:01,113
and run it somewhere else.

1085
00:51:01,113 --> 00:51:02,910
As a footnote,
someone's actually

1086
00:51:02,910 --> 00:51:05,610
running a copy of our
software in Poland.

1087
00:51:05,610 --> 00:51:07,710
And there's another copy
of the software running

1088
00:51:07,710 --> 00:51:09,475
in a server in the US, also.

1089
00:51:09,475 --> 00:51:10,600
So we aren't the only site.

1090
00:51:10,600 --> 00:51:16,380
Phil, it sounds like
this situation is rapidly

1091
00:51:16,380 --> 00:51:19,240
spinning out of control
from your perspective.

1092
00:51:19,240 --> 00:51:22,020
I mean, we have the
clear proliferation

1093
00:51:22,020 --> 00:51:25,140
of these remailers, other
kinds of encryption technology,

1094
00:51:25,140 --> 00:51:27,120
as you well know.

1095
00:51:27,120 --> 00:51:31,740
And I hear in what you
say, a subtle message which

1096
00:51:31,740 --> 00:51:37,230
says that it's really
incumbent on the technology

1097
00:51:37,230 --> 00:51:40,080
community, the internet
community, the researchers

1098
00:51:40,080 --> 00:51:43,680
here at LCS and
other places, to try

1099
00:51:43,680 --> 00:51:47,460
to direct the development
of this technology in a way

1100
00:51:47,460 --> 00:51:52,210
that, as you say, would have a
little bit more balance to it.

1101
00:51:52,210 --> 00:51:56,370
And I'm going to try
to pin you down and see

1102
00:51:56,370 --> 00:51:59,040
if you can say, what would
that mean in your eyes?

1103
00:51:59,040 --> 00:52:00,660
How would how would
the world look?

1104
00:52:00,660 --> 00:52:03,290


1105
00:52:03,290 --> 00:52:05,615
I don't know if you
can pin me down.

1106
00:52:05,615 --> 00:52:07,700
[LAUGHS] We'll see.

1107
00:52:07,700 --> 00:52:11,900
I'm a highly-trained lawyer.

1108
00:52:11,900 --> 00:52:15,050
First, I'll give an
uncharacteristically brief

1109
00:52:15,050 --> 00:52:17,060
response to what
Frans was saying,

1110
00:52:17,060 --> 00:52:19,820
and that's, Frans is right.

1111
00:52:19,820 --> 00:52:24,800
The globalization causes
us severe problems.

1112
00:52:24,800 --> 00:52:26,790
One thing computer
hackers do, for example,

1113
00:52:26,790 --> 00:52:29,452
is they break into a
series of computers

1114
00:52:29,452 --> 00:52:31,160
before they get to
their eventual target,

1115
00:52:31,160 --> 00:52:33,410
and they intentionally
across international borders

1116
00:52:33,410 --> 00:52:35,450
to make it more difficult
for law enforcement.

1117
00:52:35,450 --> 00:52:37,940
You do the same thing
with anonymous remailers.

1118
00:52:37,940 --> 00:52:40,100
You can go through a chain
of anonymous remailers

1119
00:52:40,100 --> 00:52:42,740
crossing borders, so that
if we want to try and trace

1120
00:52:42,740 --> 00:52:47,030
the communication we've got
to work with 8, or 10, or 20

1121
00:52:47,030 --> 00:52:49,910
odd countries in order
to try and trace it.

1122
00:52:49,910 --> 00:52:54,350
That is becoming a more and
more difficult burden for us

1123
00:52:54,350 --> 00:52:57,500
because it's very rare that we
deal with a computer intrusion

1124
00:52:57,500 --> 00:53:01,410
case, for example, that doesn't
have an international element.

1125
00:53:01,410 --> 00:53:03,980
But it may not be
an insuperable--

1126
00:53:03,980 --> 00:53:05,930
excuse me-- an
insuperable burden

1127
00:53:05,930 --> 00:53:09,000
if people keep the
relevant records

1128
00:53:09,000 --> 00:53:11,730
and are able to assist us.

1129
00:53:11,730 --> 00:53:15,000
The reason-- now, moving on to
your not pin me down question--

1130
00:53:15,000 --> 00:53:19,320
I cannot give you an absolute
prescription on what people

1131
00:53:19,320 --> 00:53:21,480
should do now or
shouldn't do now.

1132
00:53:21,480 --> 00:53:26,460
I think, like apparently the
MIT anonymous remailer has done,

1133
00:53:26,460 --> 00:53:29,730
they need to keep into
account what the risks are.

1134
00:53:29,730 --> 00:53:31,770
And they need to
assist law enforcement

1135
00:53:31,770 --> 00:53:34,110
when law enforcement
comes knocking

1136
00:53:34,110 --> 00:53:37,740
at the door with
appropriate legal process.

1137
00:53:37,740 --> 00:53:39,240
As I said before,
there shouldn't

1138
00:53:39,240 --> 00:53:41,760
be a first resort to law.

1139
00:53:41,760 --> 00:53:44,190
There should generally
be a first resort

1140
00:53:44,190 --> 00:53:46,500
to technology and policy.

1141
00:53:46,500 --> 00:53:48,780
And I can't give
you a prescription

1142
00:53:48,780 --> 00:53:51,630
because over the years,
and in internet time,

1143
00:53:51,630 --> 00:53:54,360
over the hours or
days, I don't know what

1144
00:53:54,360 --> 00:53:55,740
technology is going to bring.

1145
00:53:55,740 --> 00:53:57,840
I don't know how this
is going to develop.

1146
00:53:57,840 --> 00:54:01,140
This all may be moot
because we may have--

1147
00:54:01,140 --> 00:54:06,870
in six months, we may
have a universal anonymous

1148
00:54:06,870 --> 00:54:08,983
infrastructure that's going
to prevent the tracing

1149
00:54:08,983 --> 00:54:09,900
of any communications.

1150
00:54:09,900 --> 00:54:12,120
That's going to be a
very different world.

1151
00:54:12,120 --> 00:54:16,260
And that might cause Congress
to consider legal solutions.

1152
00:54:16,260 --> 00:54:20,670
But you have to look
at each step of the way

1153
00:54:20,670 --> 00:54:23,130
and see where you are, what
technology can do for you,

1154
00:54:23,130 --> 00:54:25,320
what policy can do
for you, and balance

1155
00:54:25,320 --> 00:54:28,280
the harms at that point.

1156
00:54:28,280 --> 00:54:29,340
So--

1157
00:54:29,340 --> 00:54:31,090
And I'm not going to
tell you to shut down

1158
00:54:31,090 --> 00:54:32,048
the anonymous remailer.

1159
00:54:32,048 --> 00:54:33,400
No, no, no, I know you're not.

1160
00:54:33,400 --> 00:54:38,650
But let me take a little
straw poll, because clearly

1161
00:54:38,650 --> 00:54:41,380
no one's going to say
shut it down, Michael.

1162
00:54:41,380 --> 00:54:41,980
Can we go now?

1163
00:54:41,980 --> 00:54:43,647
You should give them
a chance to say it.

1164
00:54:43,647 --> 00:54:45,730
I defend their free
speech right to say that.

1165
00:54:45,730 --> 00:54:46,390
[INAUDIBLE]

1166
00:54:46,390 --> 00:54:49,090
Of course you do,
of course you do.

1167
00:54:49,090 --> 00:54:53,770
But I don't think anybody
is even inclined, even Phil.

1168
00:54:53,770 --> 00:54:59,170
Now, let's see if we could
get to a closer question.

1169
00:54:59,170 --> 00:55:03,280
Should the policy
fact that you read

1170
00:55:03,280 --> 00:55:07,660
from, what I understand to
be, more or less, the rules

1171
00:55:07,660 --> 00:55:14,170
that the remailer operators have
agreed to guide themselves by,

1172
00:55:14,170 --> 00:55:19,570
should they say that they
will not, if they can help it,

1173
00:55:19,570 --> 00:55:24,640
or that or the remailer
should not be used in a chain

1174
00:55:24,640 --> 00:55:27,670
with a remailer in a
country to which US law

1175
00:55:27,670 --> 00:55:30,925
enforcement does not have access
through valid legal process?

1176
00:55:30,925 --> 00:55:34,040


1177
00:55:34,040 --> 00:55:37,440
In other words, let's try to
accommodate-- oh I stumped

1178
00:55:37,440 --> 00:55:37,940
Phil.

1179
00:55:37,940 --> 00:55:38,720
That's great.

1180
00:55:38,720 --> 00:55:39,570
In other words--

1181
00:55:39,570 --> 00:55:41,850
This is the way to duck
the question [INAUDIBLE]..

1182
00:55:41,850 --> 00:55:45,020
In other words, what
we've understood here

1183
00:55:45,020 --> 00:55:50,150
is that if people use
the remailer carelessly,

1184
00:55:50,150 --> 00:55:52,970
it's no problem for
law enforcement.

1185
00:55:52,970 --> 00:55:55,910
If people use the
remailer very cleverly,

1186
00:55:55,910 --> 00:55:57,830
and get good legal
advice, and figure out

1187
00:55:57,830 --> 00:56:00,600
that they need to send their
mail through a remailer chain,

1188
00:56:00,600 --> 00:56:03,080
including some country
that has no wiretapping law

1189
00:56:03,080 --> 00:56:11,330
or won't answer US court orders
or US requests for assistance,

1190
00:56:11,330 --> 00:56:16,640
then they've got a pretty
good chance of evading access

1191
00:56:16,640 --> 00:56:17,870
by US law enforcement.

1192
00:56:17,870 --> 00:56:20,600
Should these guys,
out of the goodness

1193
00:56:20,600 --> 00:56:25,670
of their hearts,
being good netizens,

1194
00:56:25,670 --> 00:56:29,390
say that they will try to
prevent that and encourage

1195
00:56:29,390 --> 00:56:34,127
people who use their service
to abide by that policy?

1196
00:56:34,127 --> 00:56:35,210
But just a technical note.

1197
00:56:35,210 --> 00:56:39,020
I mean, we would potentially
also need to keep logs for--

1198
00:56:39,020 --> 00:56:40,500
I understand you
can't enforce it.

1199
00:56:40,500 --> 00:56:41,875
I'm not asking
you to enforce it.

1200
00:56:41,875 --> 00:56:43,850
You don't want to force
any of the policies

1201
00:56:43,850 --> 00:56:46,280
in your policy fact, now.

1202
00:56:46,280 --> 00:56:48,050
All you do is you say,
here are the rules

1203
00:56:48,050 --> 00:56:51,990
for being a good citizen
using our remailer.

1204
00:56:51,990 --> 00:56:55,070
I'm proposing to
add another rule.

1205
00:56:55,070 --> 00:56:57,050
Nadine?

1206
00:56:57,050 --> 00:57:00,620
I suppose there's no harm in
encouraging people to do it,

1207
00:57:00,620 --> 00:57:04,510
as long as it's not enforceable
and there are no sanctions.

1208
00:57:04,510 --> 00:57:06,980
[LAUGHTER]

1209
00:57:06,980 --> 00:57:08,630
I defend their
right to encourage

1210
00:57:08,630 --> 00:57:10,770
as a form of advocacy.

1211
00:57:10,770 --> 00:57:13,350
But again, one point
where we agree--

1212
00:57:13,350 --> 00:57:16,820
and I think that's partly why
Phil can't be pinned down--

1213
00:57:16,820 --> 00:57:19,550
is we don't want
policy to be shaped

1214
00:57:19,550 --> 00:57:23,120
by the worst-case scenario,
because maybe the person who

1215
00:57:23,120 --> 00:57:25,880
wants that kind of
untraceable rooting

1216
00:57:25,880 --> 00:57:29,000
has what we would all agree
is not only legitimate,

1217
00:57:29,000 --> 00:57:32,590
but indeed, even a compelling
reason for doing it.

1218
00:57:32,590 --> 00:57:35,090
In fact, it may even be
a law enforcement reason.

1219
00:57:35,090 --> 00:57:36,860
Maybe this is
somebody who is going

1220
00:57:36,860 --> 00:57:41,450
to provide some tips about a
crime that's been committed,

1221
00:57:41,450 --> 00:57:45,920
but fears for his or her safety
if the identity comes to light.

1222
00:57:45,920 --> 00:57:47,900
So it may actually
be an informant

1223
00:57:47,900 --> 00:57:51,335
who's trying to help Phil
crack one of his cases open.

1224
00:57:51,335 --> 00:57:52,730
[LAUGHTER]

1225
00:57:52,730 --> 00:57:56,400
Let me see if I can answer
the question this way.

1226
00:57:56,400 --> 00:57:59,900
What we are approaching
now is perilously close

1227
00:57:59,900 --> 00:58:03,140
to a lose-lose situation
on the internet,

1228
00:58:03,140 --> 00:58:06,080
generally, and not with respect
to this anonymous remailer,

1229
00:58:06,080 --> 00:58:11,570
where legitimate citizens who
are out there surfing the web

1230
00:58:11,570 --> 00:58:13,910
have little, if no
privacy, because they're

1231
00:58:13,910 --> 00:58:17,300
tracked by commercial
interests, et cetera.

1232
00:58:17,300 --> 00:58:20,750
But people who want to
engage in criminal activity

1233
00:58:20,750 --> 00:58:23,360
can obtain nearly
perfect privacy

1234
00:58:23,360 --> 00:58:27,050
by using things such as
anonymous web browsing services

1235
00:58:27,050 --> 00:58:28,910
and anonymous emailers.

1236
00:58:28,910 --> 00:58:32,020
Essentially, we're
reaching a level

1237
00:58:32,020 --> 00:58:36,340
where just the
cognoscenti have privacy,

1238
00:58:36,340 --> 00:58:38,980
and criminals who
have motive can

1239
00:58:38,980 --> 00:58:41,170
become part of the cognoscenti.

1240
00:58:41,170 --> 00:58:46,567
And I don't think that situation
is advantageous for anyone.

1241
00:58:46,567 --> 00:58:49,561
[APPLAUSE]

1242
00:58:49,561 --> 00:58:51,560


1243
00:58:51,560 --> 00:58:54,390
What you can do
about that, Frans?

1244
00:58:54,390 --> 00:58:57,463
Well, it seems to me that good
criminals and clever criminals

1245
00:58:57,463 --> 00:58:58,880
have access to the
same technology

1246
00:58:58,880 --> 00:58:59,880
that everybody else has.

1247
00:58:59,880 --> 00:59:02,630
And if they have access
to good crypto stuff,

1248
00:59:02,630 --> 00:59:08,100
they can commit and communicate
in any degree of security

1249
00:59:08,100 --> 00:59:08,600
they like.

1250
00:59:08,600 --> 00:59:15,350
And nym.alias.net or any of the
other services, I don't think,

1251
00:59:15,350 --> 00:59:19,340
offer any additional benefit
than existing technologies

1252
00:59:19,340 --> 00:59:21,170
for committing crimes.

1253
00:59:21,170 --> 00:59:24,650
And I don't see what any
additional benefit a criminal

1254
00:59:24,650 --> 00:59:25,550
would get.

1255
00:59:25,550 --> 00:59:27,350
If you're going to
go down that line,

1256
00:59:27,350 --> 00:59:28,880
I think the debate
really is going

1257
00:59:28,880 --> 00:59:32,510
to go about whether people
should be allowed to use crypto

1258
00:59:32,510 --> 00:59:33,590
technology or not.

1259
00:59:33,590 --> 00:59:37,850
And what we're providing, or
what nym.alias.net provides

1260
00:59:37,850 --> 00:59:42,210
is basically one
application of cryptology.

1261
00:59:42,210 --> 00:59:42,710
David?

1262
00:59:42,710 --> 00:59:45,830
If I could add,
first of all, a lot

1263
00:59:45,830 --> 00:59:48,290
of the building
blocks for privacy

1264
00:59:48,290 --> 00:59:51,380
are very, very basic things
that are easy to obtain.

1265
00:59:51,380 --> 00:59:54,600
So to take a particularly,
maybe oversimple example,

1266
00:59:54,600 --> 00:59:58,113
if you have two criminals who
want to plot some terrorist act

1267
00:59:58,113 --> 01:00:00,530
and they don't want you to
know that they're communicating

1268
01:00:00,530 --> 01:00:03,250
with each other, they could
each take out classified ads

1269
01:00:03,250 --> 01:00:05,750
in The New York Times every
day, or do the equivalent, which

1270
01:00:05,750 --> 01:00:08,810
is post to this newsgroup
called alt.anonymous.messages,

1271
01:00:08,810 --> 01:00:11,270
where they're just all these
encrypted messages going by,

1272
01:00:11,270 --> 01:00:13,635
and presumably people are
downloading the ones that

1273
01:00:13,635 --> 01:00:16,010
are destined to them, but you
don't know where it's going

1274
01:00:16,010 --> 01:00:18,590
or who's reading what.

1275
01:00:18,590 --> 01:00:22,070
Definitely one of the
goals of nym.alias.net

1276
01:00:22,070 --> 01:00:24,780
is to make this
technology more available.

1277
01:00:24,780 --> 01:00:27,750
And we've done it in at least
sort of a one-sided way.

1278
01:00:27,750 --> 01:00:30,800
So one thing that people were
particularly appreciative

1279
01:00:30,800 --> 01:00:33,380
of in the surveys we sent
out is that, for example,

1280
01:00:33,380 --> 01:00:36,565
this humanitarian
aid worker said

1281
01:00:36,565 --> 01:00:38,690
that there was no way his
or her friends were going

1282
01:00:38,690 --> 01:00:40,670
to learn how to use PGP,
and encryption software,

1283
01:00:40,670 --> 01:00:41,460
and all this stuff.

1284
01:00:41,460 --> 01:00:43,918
But this person was willing to
put in the time to learn it.

1285
01:00:43,918 --> 01:00:46,850
And then, this person could use
nym.alias.net to communicate

1286
01:00:46,850 --> 01:00:47,630
with other people.

1287
01:00:47,630 --> 01:00:50,360
And the nym just looked like
an ordinary email address.

1288
01:00:50,360 --> 01:00:52,797
So that at least now you
can communicate with people.

1289
01:00:52,797 --> 01:00:54,380
If you need privacy,
you can learn it.

1290
01:00:54,380 --> 01:00:58,610
And you don't need to make all
your friends learn this stuff.

1291
01:00:58,610 --> 01:01:00,410
And again, there's
a lot of work to be

1292
01:01:00,410 --> 01:01:02,602
done in making this
stuff easier to use.

1293
01:01:02,602 --> 01:01:04,310
And people have actually
written software

1294
01:01:04,310 --> 01:01:05,840
to help make it easier to use.

1295
01:01:05,840 --> 01:01:08,370
But I absolutely agree that
there's a lot to be done.

1296
01:01:08,370 --> 01:01:12,950
And I hope that things happen to
make privacy easier for people

1297
01:01:12,950 --> 01:01:14,900
who are not computer literate.

1298
01:01:14,900 --> 01:01:17,320
I definitely think that's
an important problem.

1299
01:01:17,320 --> 01:01:20,380
Well, David, I think you're
bringing us to a close here.

1300
01:01:20,380 --> 01:01:22,900
I think that we've
learned, on the one hand,

1301
01:01:22,900 --> 01:01:26,020
that no one's ready to
shut this remailer down.

1302
01:01:26,020 --> 01:01:30,130
I also think we've learned
that the lose-lose scenario

1303
01:01:30,130 --> 01:01:34,960
that Phil describes where
privacy is inaccessible to most

1304
01:01:34,960 --> 01:01:40,540
and, in fact, used harmfully
by some is also not a scenario

1305
01:01:40,540 --> 01:01:43,920
judging by the Gallup
poll of this audience,

1306
01:01:43,920 --> 01:01:45,940
that anyone wants to encourage.

1307
01:01:45,940 --> 01:01:47,890
In the spirit of
this conference,

1308
01:01:47,890 --> 01:01:50,860
in the spirit of looking
forward 35 years,

1309
01:01:50,860 --> 01:01:53,090
creating the future
world of information,

1310
01:01:53,090 --> 01:01:54,790
I just want to ask all of you--

1311
01:01:54,790 --> 01:01:57,493
David gave his closing
hopes, I think.

1312
01:01:57,493 --> 01:01:59,410
I want to ask all of
you, starting with Frans,

1313
01:01:59,410 --> 01:02:03,160
to just give us a
sense of how you'd

1314
01:02:03,160 --> 01:02:07,720
like the techniques of
anonymity to look on the net

1315
01:02:07,720 --> 01:02:10,850
as we go forward.

1316
01:02:10,850 --> 01:02:15,155
I guess the two--

1317
01:02:15,155 --> 01:02:16,280
one way [INAUDIBLE] common.

1318
01:02:16,280 --> 01:02:20,600
Basically, I think if we like
to see the technology develop

1319
01:02:20,600 --> 01:02:24,290
in the direction where more
and more people can get access

1320
01:02:24,290 --> 01:02:27,380
to technology easily so that
their privacy is protected.

1321
01:02:27,380 --> 01:02:31,730
And as Nadine pointed out, it's
sort of a wild wild west there

1322
01:02:31,730 --> 01:02:36,640
on the internet, and privacy
is really ill-protected,

1323
01:02:36,640 --> 01:02:38,630
and there's a lot of
commercial efforts going

1324
01:02:38,630 --> 01:02:40,880
on trying to figure
out who somebody is

1325
01:02:40,880 --> 01:02:42,410
and trying to record things.

1326
01:02:42,410 --> 01:02:45,060
And as was pointed
out in the talk,

1327
01:02:45,060 --> 01:02:47,900
and, in fact, was pointed
out in the haystack talk,

1328
01:02:47,900 --> 01:02:51,740
it's basically you say something
now, or do something now,

1329
01:02:51,740 --> 01:02:53,690
and 10 years later,
and you can still

1330
01:02:53,690 --> 01:02:57,300
be able to find it and
exactly discover what you did.

1331
01:02:57,300 --> 01:03:00,110
And it also came out, I think,
in Ron's talk very clearly.

1332
01:03:00,110 --> 01:03:05,670
So figuring out technology that
will protect privacy of users

1333
01:03:05,670 --> 01:03:08,170
in the internet, I think, is
going to be a crucial question,

1334
01:03:08,170 --> 01:03:10,760
and I would see that
like to to happen.

1335
01:03:10,760 --> 01:03:12,420
At the same time,
I think it's going

1336
01:03:12,420 --> 01:03:14,810
to be crucial to figure out
in the next-- whatever--

1337
01:03:14,810 --> 01:03:18,320
a couple of decades, what
this division is between law

1338
01:03:18,320 --> 01:03:20,930
enforcement, or under the
laws, and the legal laws

1339
01:03:20,930 --> 01:03:22,610
in the technology.

1340
01:03:22,610 --> 01:03:25,083
And that struck me as
a very important point,

1341
01:03:25,083 --> 01:03:26,750
because we certainly
want to be the case

1342
01:03:26,750 --> 01:03:34,462
that if somebody commits a
crime, they face prosecution.

1343
01:03:34,462 --> 01:03:36,170
Since this is the sort
of final comments,

1344
01:03:36,170 --> 01:03:38,660
I want to make and one
more other comment.

1345
01:03:38,660 --> 01:03:42,140
And actually, I want to thank
all the people that actually

1346
01:03:42,140 --> 01:03:44,770
have made it possible to
be able to run the service

1347
01:03:44,770 --> 01:03:46,020
over the last couple of years.

1348
01:03:46,020 --> 01:03:47,060
I mean, there are a lot
of people, actually, some

1349
01:03:47,060 --> 01:03:50,300
in the audience, so many
outside of here that helped us

1350
01:03:50,300 --> 01:03:52,520
in running the
service and dealing

1351
01:03:52,520 --> 01:03:55,850
with some of the problematic
issues that we faced.

1352
01:03:55,850 --> 01:03:58,847
And that's why I want to make
sure that those people get

1353
01:03:58,847 --> 01:03:59,930
the appropriate thank you.

1354
01:03:59,930 --> 01:04:03,494
And Phil, you can thank those
people, too, if you choose.

1355
01:04:03,494 --> 01:04:05,370
Please.

1356
01:04:05,370 --> 01:04:05,950
Am I next?

1357
01:04:05,950 --> 01:04:06,450
Yes.

1358
01:04:06,450 --> 01:04:07,300
OK.

1359
01:04:07,300 --> 01:04:11,020
I'll briefly return to an
analogy Nadine drew earlier,

1360
01:04:11,020 --> 01:04:13,960
because I'm not really sure
she'd like where it goes.

1361
01:04:13,960 --> 01:04:18,730
She indicated briefly
that I was a person who

1362
01:04:18,730 --> 01:04:22,940
wanted to ban the use
of the automobile.

1363
01:04:22,940 --> 01:04:25,180
But if you think
about the illogical

1364
01:04:25,180 --> 01:04:29,740
result of that analogy, we
have people use automobiles,

1365
01:04:29,740 --> 01:04:31,960
but you have to have a
driver's license to use them,

1366
01:04:31,960 --> 01:04:34,970
and you have to identify
yourself to drive the car.

1367
01:04:34,970 --> 01:04:37,270
So I think the logical
extension of that is everyone

1368
01:04:37,270 --> 01:04:38,687
can use computers
on the internet,

1369
01:04:38,687 --> 01:04:40,930
but you've got to identify
yourself to government

1370
01:04:40,930 --> 01:04:44,200
and get a license to do it.

1371
01:04:44,200 --> 01:04:47,170
I don't think anybody
here wants to go there.

1372
01:04:47,170 --> 01:04:50,245
So let's just leave
that analogy behind.

1373
01:04:50,245 --> 01:04:55,740


1374
01:04:55,740 --> 01:04:58,680
I guess the concluding
point comes from something

1375
01:04:58,680 --> 01:05:01,770
that Frans said, which is--

1376
01:05:01,770 --> 01:05:04,590
and this comes up all the time
in the encryption debate--

1377
01:05:04,590 --> 01:05:07,602
the best criminals
are going to use--

1378
01:05:07,602 --> 01:05:09,060
no matter what
government policy is

1379
01:05:09,060 --> 01:05:10,435
in encryption--
strong encryption

1380
01:05:10,435 --> 01:05:12,360
that doesn't allow
for government access.

1381
01:05:12,360 --> 01:05:14,070
The best criminals
on the internet

1382
01:05:14,070 --> 01:05:17,340
are always going to be
able to obtain anonymity.

1383
01:05:17,340 --> 01:05:21,210
In designing systems,
in designing procedures,

1384
01:05:21,210 --> 01:05:24,060
for example, building in
things like digital signatures

1385
01:05:24,060 --> 01:05:26,640
into products that will
help law enforcement

1386
01:05:26,640 --> 01:05:28,710
determine who's engaging
in communications,

1387
01:05:28,710 --> 01:05:30,750
if it's a part of a standard.

1388
01:05:30,750 --> 01:05:35,790
In designing products, and in
designing policies and systems,

1389
01:05:35,790 --> 01:05:38,220
if we can do that
so that you can

1390
01:05:38,220 --> 01:05:41,850
help law enforcement catch
the stupid criminals,

1391
01:05:41,850 --> 01:05:43,057
that's a big help.

1392
01:05:43,057 --> 01:05:46,470
[APPLAUSE]

1393
01:05:46,470 --> 01:05:51,810
Nadine, the fate of the smart
criminals are in your hands.

1394
01:05:51,810 --> 01:05:54,810
I don't want to get too much
off into the driving analogy,

1395
01:05:54,810 --> 01:05:56,970
but I want to assure
everybody out there

1396
01:05:56,970 --> 01:06:01,830
that the ACLU is constantly,
and to some degree with success,

1397
01:06:01,830 --> 01:06:06,330
challenging violations of
privacy that are now committed

1398
01:06:06,330 --> 01:06:08,760
in the name of driving.

1399
01:06:08,760 --> 01:06:13,080
In fact, here in
Massachusetts, we recently

1400
01:06:13,080 --> 01:06:17,190
won the right of drivers to keep
their Social Security numbers

1401
01:06:17,190 --> 01:06:19,350
off those licenses, because
that's, as you know,

1402
01:06:19,350 --> 01:06:22,490
a major way to
invade data privacy.

1403
01:06:22,490 --> 01:06:26,740
But when Danny asked us
to look ahead 35 years,

1404
01:06:26,740 --> 01:06:28,440
I did a little bit
of calculation here

1405
01:06:28,440 --> 01:06:33,390
and I thought, where were we
35 years ago, which was 1964.

1406
01:06:33,390 --> 01:06:36,750
In 1964, the United
States Supreme Court

1407
01:06:36,750 --> 01:06:41,130
was still ruling that
wiretapping was not

1408
01:06:41,130 --> 01:06:44,550
a search and seizure that was
subject to the Fourth Amendment

1409
01:06:44,550 --> 01:06:45,060
at all.

1410
01:06:45,060 --> 01:06:48,540
That was a decision
they made back in 1927

1411
01:06:48,540 --> 01:06:53,430
over an eloquent dissent by
Justice Louis Brandeis who

1412
01:06:53,430 --> 01:06:58,110
issued ringing words that are
still uncannily true today.

1413
01:06:58,110 --> 01:06:59,610
And unfortunately,
I'm going to have

1414
01:06:59,610 --> 01:07:02,430
to paraphrase because I didn't
bring the quote with me.

1415
01:07:02,430 --> 01:07:06,780
But basically the concept
was that government is always

1416
01:07:06,780 --> 01:07:09,630
going to come up with
new technologies that

1417
01:07:09,630 --> 01:07:13,980
can be used to invade
individual freedom even more

1418
01:07:13,980 --> 01:07:18,090
effectively than the means that
were available that prompted

1419
01:07:18,090 --> 01:07:20,760
the framers to write the
Fourth Amendment, the First

1420
01:07:20,760 --> 01:07:24,810
Amendment, other individual
freedoms into our Constitution.

1421
01:07:24,810 --> 01:07:28,590
And we cannot just throw
those freedoms to the wind

1422
01:07:28,590 --> 01:07:32,730
because of the initial
temptation to be panicked about

1423
01:07:32,730 --> 01:07:36,630
the potential abuses
of the new technology.

1424
01:07:36,630 --> 01:07:38,790
But we've repeatedly
gone through this cycle,

1425
01:07:38,790 --> 01:07:41,730
not only in American history,
but around the world when

1426
01:07:41,730 --> 01:07:43,800
the printing press
was first invented,

1427
01:07:43,800 --> 01:07:45,870
panic on the part
of governments.

1428
01:07:45,870 --> 01:07:49,050
We've got to license it, prior
restraint, all of this danger

1429
01:07:49,050 --> 01:07:51,390
can be used by criminals,
it can corrupt children,

1430
01:07:51,390 --> 01:07:53,310
the same litany
that we subsequently

1431
01:07:53,310 --> 01:07:56,220
heard with every new
technology, and now we're

1432
01:07:56,220 --> 01:07:57,420
hearing with the internet.

1433
01:07:57,420 --> 01:08:00,360
But after a while,
the panic subsides.

1434
01:08:00,360 --> 01:08:03,540
People understand that it's
more alike than it's dissimilar,

1435
01:08:03,540 --> 01:08:07,440
that the positive uses far
outweigh the dangerous uses,

1436
01:08:07,440 --> 01:08:10,690
that alternative means can be
found to deal with the danger.

1437
01:08:10,690 --> 01:08:14,280
So I hope that, and
believe strongly,

1438
01:08:14,280 --> 01:08:18,450
that 35 years from now, the
panic about this supposed

1439
01:08:18,450 --> 01:08:22,920
unique dangers of the internet
will seem as antiquated

1440
01:08:22,920 --> 01:08:27,810
as the panic that first withdrew
constitutional protection

1441
01:08:27,810 --> 01:08:30,880
from earlier technologies.

1442
01:08:30,880 --> 01:08:32,040
Thank you, Nadine.

1443
01:08:32,040 --> 01:08:33,779
I think that the
panel has really

1444
01:08:33,779 --> 01:08:39,420
done us a tremendous service,
because not only have you

1445
01:08:39,420 --> 01:08:42,750
really helped us to think
through the issues about today

1446
01:08:42,750 --> 01:08:45,100
and what we ought
to be doing today,

1447
01:08:45,100 --> 01:08:46,950
but I think that
you've helped us

1448
01:08:46,950 --> 01:08:49,950
really to frame what we all
have to be thinking about

1449
01:08:49,950 --> 01:08:52,920
as we try to put privacy
and other important values

1450
01:08:52,920 --> 01:08:55,020
into oxygen going forward.

1451
01:08:55,020 --> 01:08:59,229
With that, back to Bob Metcalf,
our able master of ceremonies.

1452
01:08:59,229 --> 01:08:59,939
Thanks very much.

1453
01:08:59,939 --> 01:09:00,500
[APPLAUSE]

1454
01:09:00,500 --> 01:09:01,150
Thank you, all.

1455
01:09:01,150 --> 01:09:07,260


1456
01:09:07,260 --> 01:09:10,020
This marks the end of the
presentations for the day.

1457
01:09:10,020 --> 01:09:11,600
Don't forget that
tomorrow morning

1458
01:09:11,600 --> 01:09:16,282
we start sharply at 8:30 at
the nearby Johnson Center.

1459
01:09:16,282 --> 01:09:17,990
And to help you find
it tomorrow morning,

1460
01:09:17,990 --> 01:09:21,793
we're going to go there now
just to show you where it is.

1461
01:09:21,793 --> 01:09:23,210
Incidentally, I
want to remind you

1462
01:09:23,210 --> 01:09:27,529
that we have a special security
event tomorrow in connection

1463
01:09:27,529 --> 01:09:29,779
with the visit of
the Chinese Premier.

1464
01:09:29,779 --> 01:09:31,850
So you could come
early if you wanted.

1465
01:09:31,850 --> 01:09:34,220
It would be OK to come
early, rather than try

1466
01:09:34,220 --> 01:09:37,580
to think you're going to get
in just at the last minute.

1467
01:09:37,580 --> 01:09:39,229
And now, we're
going to move over

1468
01:09:39,229 --> 01:09:42,590
to the Johnson Center for
a reception and for demos.

1469
01:09:42,590 --> 01:09:44,720
There's a bunch of demos
waiting for you now

1470
01:09:44,720 --> 01:09:46,580
at the Johnson Center.

1471
01:09:46,580 --> 01:09:47,960
Good evening.

1472
01:09:47,960 --> 01:09:51,010
[APPLAUSE]

1473
01:09:51,010 --> 01:09:59,000