IGF-USA 2017 - 24 July 2017 - Carlos A Bulgheroni Room CSIS,
Washington DC

1 - Nationalism, Disinformation and Free Expression in the Age of
the Internet
https://www.igf-usa.org/nationalism-disinformation-and-free-
expression/

KEYNOTE - Vint Cerf,  Chief Internet Evangelist for Google

=========

SHANE TEWS: We're going to open with Vint Cerf, which is always a
pleasure. If you ever have a moment where you have 15 minutes, there
is a great YouTube video, where Vint Cerf was interviewed for this
conference two years ago, with Steve Crocker, and it's just one of
my favorites. I have to admit, I've watched it several times,
because he just goes through all of these moments where he's...
Well, we were thinking about this, and we tried that and it worked
and it seems like it's still working and it's great. We all get the
pleasure of his smart ideas on an everyday basis, and I can't thank
him enough. He kind of got this whole Internet party started for us,
and all the things we get to do, because there are gentlemen like
him who were smart enough to realize that there were things we could
be doing with the web technology beyond just using it for its
limited purposes, back in the '60s and early '70s. So Vint, if you
wouldn't mind coming up and just getting us started here, come on
up.

VINT CERF: Well, good morning, everyone. It's a real pleasure to
participate in IGF-USA. And I wanted to make a couple of
observations about the regional and national IGF meetings. I think
they are incredibly important, especially given that this phenomenon
is sort of bottom-up. These things did not happen in consequence of
the UN-sponsored IGF. They happened because people thought they
needed to have conversations locally about issues arising. So I want
to encourage you to participate in these things in the future.
Because even if the international IGF were to evaporate for some
reason, I think the aggregate of the national and regional IGFs
could combine together to support further and continued
international multistakeholder meetings. I have the convenience of
being able to stand up here and state the obvious, because I'm the
first speaker, and that way, I will say all the things that you
already know, and all the other speakers are likely to say. But I
can get away with it, because I'm the first speaker.

So let me tell you how this current situation, with the Internet,
looks to somebody who's been around and engaged from its very
beginning. We thought it was important to have an open and free
Internet, freely accessible Internet, and as you know, Bob and I
gave away the design deliberately, published all the specifics,
encouraged the Internet Engineering Task force and the Internet
Architecting board to continue in a very open and multistakeholder
way. The protocols were very open. Technically speaking, people
could add new protocols if they wanted to, both horizontally and
vertically, and of course, that is exactly what has happened, when
the world wide web came along. Tim Berners-Lee layered HTTP on top
of TCP/IP. All of this led to what many of us call permissionless
innovation. All of which was very satisfying to me, watching this
grow in a very organic way.

There's only one small little detail that had not penetrated my
thinking in the early stages, and that's: What happens when the
general public gets access? And this was something I promoted
strongly, especially in the late 1980s. The problem is that the
general public is the general public. And it covers everybody,
including the bad actors, who in fact do not have your best
interests at heart. And so now, having created this giant engine,
which gave everyone the freedom to speak, now all the bad guys also
had freedom to speak, and they spoke in some fairly harmful ways. On
the technical side, these bad actors could speak malware. And you
know what DDoS attacks are. They're amplified DDoS attacks. These
are the ones that make use of the domain name system. They use fake
source addresses to make a query to the DNS, and the response goes
to whatever the fake address was, and the fake address is the target
of the amplified DDoS attack. So the DNS, which is part of the
infrastructure of the Internet, is used to amplify an attack against
a particular target.

Now, there are technical responses to many of these things, although
frankly not all the ISPs have implemented those technical responses.
For the geeks in the crowd, BCP-38 is designed to inhibit the
injection of packets that have false source addresses in them, but
it hasn't been widely implemented. If there are any ISPs in the room
who haven't implemented it, shame on you. But I can't help but
observe that Twitter is another example of an amplified DDoS attack.
Because if you tweet something, if it gets retweeted by a collection
of people, that's an amplification. So in some sense, we are
recurring in the social networking space, in the way that -- and
discovering the same kinds of harmful problems that showed up in the
technical space.

Now, I'm not gonna go any further on the technical responses to many
of these problems, except to say that one of the hardest problems is
that programmers don't know how to write software that doesn't have
bugs. And that's why malware works. And so those of us who care
greatly about the research into programming really wish that we had
better tools to keep ourselves from making mistakes that can be
exploited later on. But I have to tell you that we are in our
infancy, when it comes to that. We've had 80-plus years of
programming in one way or another, and have failed miserably to find
ways to inhibit some of the stupid bugs that we still put into our
programs.

But I want to move over now to what I think is the more critical
question for this convening. By the way, the summary of the meetings
and everything is fabulous. I mean, you have just an absolutely
amazing array of people here, and I'm kicking myself, because I have
to leave at 10:30 to get to the West Coast, to work on the
interplanetary Internet, at the Jet Propulsion Lab, and I know we're
very concerned about Martian porn, because we don't know how to
recognize it, so we don't know how to filter it out.

So I want to switch over to the social networking side of things,
because that is in some sense part of the problem that we're
grappling with today. Many of the themes that you'll be dealing
with. I think that we can see that social networks produce a kind of
amplification. There is a bubble reinforcement effect. Sometimes
this is called bias, reinforcement bias, where you take a piece of
information that matches your view of things, and by seeing it
repeated in a social networking environment, it reinforces your
belief that that's the correct thing and anything else is not. So we
have this bubble effect, which is starting to show up in the
Internet. And it's funny in a way, because on the surface, it looked
like it would be a helpful thing to steer you in the direction of
things you care about and are interested in. But the trouble is that
it creates this amplification and reinforcement effect which
isolates people from information that would have, perhaps, persuaded
them to consider alternative views. So we have that problem to worry
about.

And there have been some attempts in the social networking space --
Facebook and others -- to try to -- even at Google -- to try to
expose people to things that they didn't necessarily appear to want
to get access to. The Arab Spring was a very good, concrete example
of the way in which people have decided to use social networking
techniques. Facebook in particular. And so what this told us, I
think, is that people will figure out how to use our technologies in
ways that we might not have anticipated. You can see the reaction of
authoritarian regimes to this kind of use of online facilities. It's
scary to think that a population that you are trying to control has
the ability to bypass government limitations, and to coordinate and
to organize. If you were in the Chinese government, for example, and
you were watching the Arab Spring happen, I think this would scare
you. And those of you who know Chinese history better than I do may
recall that almost all of the major regime changes in China, as you
move from one dynasty to another, were preceded by peasant
rebellions of one kind or another. And so if you were currently in
the Chinese government, high government, you would be looking at 1.2
or 1.3 billion people, worried about this particular risk.

What I find absolutely amazing is that the Chinese government has
simultaneously done two things. They have invested mightily in the
Internet infrastructure. 700-million-plus Chinese are online in
China. So there's been huge investment in implementation. Fiber
networks, online assets of various kinds. And at the same time, they
have introduced a substantial degree of control. And this is scary,
in some respects, and painful to watch, because of course many of us
hope that the Internet would inhibit exactly that kind of control.
But in fact the Chinese have demonstrated it's possible to filter a
great deal of content, to use legal methods, and I suppose some
people would say illegal methods, in order to limit what people are
able to say. You notice that there are some very big companies that
have evolved in the Chinese Internet environment. AliBaba, for
example, Baidu, WeChat. These companies are substantial in scale.
They rival, in terms of numbers of users, many of the other large
companies that you see here in the United States and elsewhere. But
they are largely under Chinese control with regard to censorship.

So what we're seeing is a kind of reaction to the openness and
freedom in the Internet. It translates into fragmentation, it
translates into some other things, which are even more disturbing
for me personally, anyway. And these are attempts by national
governments to extend in an extraterritorial way their control over
content. And so there are debates right now going on, leading up to
the European Court of Justice, which in my view may be a kind of
misnomer, that this question of the right to be forgotten should not
simply be limited to the European countries, but it should be global
in scope. We're seeing similar kinds of behavior in other countries.
Even in Canada, for example, where the debate is: Should information
that should be -- if it should be adjudicated, that this information
should be suppressed in Canada, should it be suppressed everywhere
in the world? These are not good outcomes for those of us who
believe that openness and freedom of expression is vital to a
democratic society.

Well, that leads to yet another phenomenon, which all of you are
very familiar with, especially given the most recent presidential
campaign in the United States. And that's misinformation and the so-
called "fake news". Again, this manifestation is another example of
the kind of selection bias and silo reinforcement that I mentioned
earlier. There was a Russian disinformation campaign, and they are,
as many of you know, quite skilled at this kind of propaganda. It's
not the first time that they've made use of it, although this might
be the first time they've used it heavily in an online environment.
And apparently, people were making money out of this campaign. In
Macedonia, if I remember correctly, people were paid to generate
completely ridiculous articles about Hillary Clinton or others who
were part of the presidential competition. So it's sort of ironic
that one of the poorest countries in Europe turned out to be making
income out of generating fake news. And the worst part about all
this is that this fake news was accepted in many circles, here in
the United States. And the question is: Why is that? How can this
happen? What's going on here?

Well, part of it is an uncritical audience. Or a polarized audience.
Where the fake news somehow reinforced their beliefs, even though
they may have made absolutely no logical sense whatsoever. It's this
uncriticality that really disturbs me a lot. I think that we should
be teaching children how to think critically about the information
that they get. They should ask: Where did it come from? They should
ask: Who else believes this information? What other sources are
there? Can we find a way to confirm the accuracy of the information
that we're receiving in this system? And the fact that there are a
lot of people who don't care to waste time thinking about the
information they get is very disturbing to me. It's not just the
Internet that creates this problem. It's all the other media as
well. You get information -- misinformation -- from television,
radio, movies, magazines, newspapers, the Internet. Your friends.
Your parents. There are all kinds of ways in which to get
information which is incorrect. And not thinking about it is very
disturbing.

However, you get the other side of this coin. There are some
families who have the belief that there is this authority, it rests
in the family, and any information which the family doesn't agree
with should be rejected. Even if it turns out that what the family
believes is in fact wrong. And so you get another strong biasing
effect in some parts of our society, here in the US, where certain
families will reject any notion of critical thinking, because they
consider that it undermines the authority of the family. So I find
that kind of scary.

There's another big problem, which contributes to this situation.
And that's the failing business models for journalism. In the past,
paper, newspaper in particular, turned out to be one of the cheapest
ways of reproducing large quantities of information on a regular
basis. And since everyone wanted to know what the news was, the
people who developed the notion of newspaper also put in
advertisements, because they figured... Well, they're gonna read the
news and they'll see the ads, and I can charge people for that. It
was perfect. It had a few features associated with the news cycle.
You had to get your stories done by a certain time, in order to
print the news and get the newspapers delivered. And so there
evolved out of this -- I left out classified ads, which are another
wonderful way of generating money. Now, we all know that many of
those revenue generators have evaporated and reincarnated themselves
in the online environment. So that undermined some of the business
models that led to substantial quality journalism, because the
newspapers could afford to do investigative reporting and to pay
people for that, to pay people to be onsite, all around the world,
providing content. So as those business models started to evaporate,
that created a real problem for quality journalism. It's fair to
say, by the way, that those who say... Well, you know, Google and
the Internet have destroyed the news business... I would like to
resist that conclusion a little bit. And argue that at least in our
society, we were drifting away from the newspaper as a source of
information, as radio, and especially as television came along.
People were turning to those media rather than papers, to get their
information. And there was a certain impatience in our society that
I think limited our willingness to spend time absorbing and
analyzing and evaluating new information.

If you don't mind a small anecdote, some years ago, maybe 10 years
ago or so, I was in New York, having lunch with Henry Kissinger. And
we sat down, and Dr. Kissinger said to me... I hate the Internet! I
thought... Well, lunch is over. I'm leaving now. But I said... Well,
why is that, Dr. Kissinger? And he said... Well, people are
satisfied with two-paragraph answers, and I write 700-page books.
And so I could see his logic. He also said something else, which is
not relevant to this conversation, except tangentially. He said he
was also very unhappy with the fact that his grandchildren could not
read cursive writing. They weren't being taught cursive. Which meant
that the huge collection of historically important letters that he
had were not accessible to his grandchildren. And I actually stopped
to think... My God. That's right. That kids don't see cursive very
much anymore at all. They see printed material. So as I say, that's
a tangential observation. But I think that Kissinger was right about
this satisfaction with too little information. And I think this dogs
us still today.

Well, so we have this problem of trying to reinvent journalism,
reinvent business models. In the mean time, you know, we have this
headline writing that is intended to capture eyeballs and not
necessarily intended to convey reality. And so we get headline-
grabbing ads and things like that. And loss of revenues. There is a
silver lining, however. At least, in my view. When Jeff Bezos bought
the Washington Post for an insanely small amount of money, relative
to his resources, my impression as a Washington resident for 40
years now, is that the quality of the Washington Post journalism has
increased again. And even if you don't agree with me, I think that
they are certainly in a position to put more effort into gathering
news, analyzing it, and doing more investigative reporting. And I
think that's partly a result of Jeff's willingness to underwrite
that cost.

But somewhere along the line, we are going to have to figure out how
to reincarnate business models that will support good quality
journalism, because of its essential importance in open and free
democratic societies. So I want to try to summarize a little bit
here ways in which we might respond to some of these problems. The
first one, I'll reiterate again, is early training in critical
thinking. And I really am sad to think that there are people who
would resist such an initiative. I don't understand how we could
possibly have a society which is able to evolve in the face of all
these new technologies, without being able to think critically about
what we're seeing and hearing. We clearly have to reinvent the
business of news. And I don't have really good answers for you. I
wish I did. Short of suggesting that others like Jeff should keep
buying newspapers. Maybe we should suggest that the other
billionaires from the Internet world consider acquiring all the
newspapers that are failing. On the other hand, if that becomes too
concentrated, then we have a different problem on our hands, which
we've also experienced in the past.

There's also been suggestions that somehow we should be able to
automate the process of filtering what we're calling "fake news" and
misinformation. And I've given some thought to this, and it turns
out it's not as easy as it sounds. For example, if you try to use a
kind of voting mechanism that says this news is valid and this news
is not, people who create botnets can use the power of the botnet to
upvote fake information. If they want to. And so the algorithms that
often are used to try to decide whether something is important or
something is significant get distorted by mechanisms that automate
these upvoting processes. And we saw a lot of that. I mean, the
battle of the bots could very well describe the previous
presidential campaign. And it continues to dog us today. So figuring
out how to detect that sort of thing -- I mean, I'm sure many of you
have logged onto a website and seen a little thing that says: I am
not a robot! And you're supposed to click the... Of course, it's
pretty easy to write a piece of software that sees "I am not a
robot" in the web page and clicks that little box. So we have the
problem of the software getting smarter and smarter, and in some
sense, we're defeating ourselves. I think that what would be really
important...

JOLY MACFIE: Excuse me. We lost the screen. I'm just checking... It
seems like the computer went to sleep.

VINT: Well, I don't care if the computer went to sleep. I'm more
concerned whether everybody in the audience went to sleep!

(laughter)

And you're not even allowed to bring the coffee into the room here,
which is really terrible!

So here we are. You've got this fabulous day ahead of you. To arm
wrestle with some of the problems that I've tried to outline, and
others that are shown in the program. I think what will be really
helpful and important is at the end of the day if you can collect
some thoughtful and practical ideas for combating the problems that
we're seeing -- it would be super helpful. Because I would love to
see you bring those to the Geneva meeting in December. If we think a
little bit about the value of these regional and national IGFs, it's
assembling thoughtful outcomes and bringing them to the
international meeting, and to draw attention to some of those
solutions. And of course, to compare with each other, with the other
IGFs, regional IGFs, the conclusions that we reached. But in some
sense, the fact that you're here in this room says you care about
this agenda. And I hope that some significant fraction of you will
be able to bring anything that came out of this discussion, that you
consider to be practical and implementable, to that table.

One last point. As you look at the Internet, as we see it today, you
could reasonably ask: Is this helping our societies? Or is it
harming it? And I think you would find answers on both ends of that
scale. But what I would like to ask you to do is to think more about
how we can make the Internet more useful for people. Let's make it a
more people-centered system. Something which is taking into account
solving problems for people, helping people discover each other, and
ways in which they can help each other to make life a lot better for
us and others in this world. It would be really disappointing if it
turned out that all of this, for me, 40-year effort into the
Internet, turned out to produce something which turned out to be
more harmful than beneficial. And I'm sort of relying on you in this
room to make sure that doesn't happen.

So I think I'll stop there. And thank you all very much for your
morning's attention.

(applause)

SHANE TEWS: Thank you very much.

==============