Alternatives to Censorship: Interview With Matt Stoller
As Congress once again demands that Silicon Valley crack down on speech, the
Director of Research at the American Economic Liberties Project outlines the
real problem - and better solutions
Matt Taibbi
Mar 27
Led by Chairman Frank Pallone, the House Energy and Commerce Committee
Thursday held a five-hour interrogation of Silicon Valley CEOs entitled,
"Disinformation Nation: Social Media's Role in Promoting Extremism and
Misinformation."
As Glenn Greenwald wrote yesterday, the hearing was at once agonizingly
boring and frightening to speech advocates, filled with scenes of members of
Congress demanding that monopolist companies engage in draconian crackdowns.
Again, as Greenwald pointed out, one of the craziest exchanges involved
Texas Democrat Lizzie Fletcher:
Fletcher brought up the State Department's maintenance of a list of Foreign
Terrorist Organizations. She praised the CEOs of Twitter, Facebook, and
Google, saying that "by all accounts, your platforms do a better job with
terrorist organizations, where that post is automatically removed with
keywords or phrases and those are designated by the state department."
Then she went further, chiding the firms for not doing the same
domestically. asking, "Would a federal standard for defining a domestic
terror organization similar to [Foreign Terrorist Organizations] help your
platforms better track and remove harmful content?"
At another point, Fletcher noted that material from the January 6th protests
had been taken down (for TK interviews of several of the videographers
affected, click here) and said, "I think we can all understand some of the
reasons for this." Then she complained about a lack of transparency, asking
the members, "Will you commit to sharing the removed content with Congress?"
so that they can continue their "investigation" of the incident.
Questions like Fletcher's suggest Congress wants to create a multi-tiered
informational system, one in which "data transparency" means sharing content
with Congress but not the public.
Worse, they're seeking systems of "responsible" curation that might mean
private companies like Google enforcing government-created lists of bannable
domestic organizations, which is pretty much the opposite of what the First
Amendment intended.
Under the system favored by Fletcher and others, these monopolistic firms
would target speakers as well as speech, a major departure from our current
legal framework, which focuses on speech connected to provable harm.
As detailed in an earlier article about NEC appointee Timothy Wu, these
solutions presuppose that the media landscape will remain highly
concentrated, the power of these firms just deployed in a direction more to
the liking of House members like Fletcher, Pallone, Minnesota's Angie Craig,
and New York's Alexandria Ocasio-Cortez, as well as Senators like Ed Markey
of Massachusetts. Remember this quote from Markey: "The issue isn't that the
companies before us today are taking too many posts down. The issue is that
they're leaving too many dangerous posts up."
Glenn Greenwald @ggreenwald
Remember: the last time Congress summed social media CEOs to be
interrogated, Sen. @EdMarkey left no doubt about what the demand of
Democrats is when doing this: we want you to censor more (and obviously, the
content they want censored is from their political adversaries):
February 20th 2021
298 Retweets781 Likes
These ideas are infected by the same fundamental reasoning error that drove
the Hill's previous drive for tech censorship in the Russian misinformation
panic. Do countries like Russia (and Saudi Arabia, Israel, the United Arab
Emirates, China, Venezuela, and others) promote division, misinformation,
and the dreaded "societal discord" in the United State? Sure. Of course.
But the sum total of the divisive efforts of those other countries makes up
at most a tiny fraction of the divisive content we ourselves produce in the
United States, as an intentional component of our commercial media system,
which uses advanced analytics and engagement strategies to get us upset with
each other.
As Matt Stoller, Director of Research at the American Economic Liberties
Project puts it, describing how companies like Facebook make money:
It's like if you were in a bar and there was a guy in the corner that was
constantly egging people onto getting into fights, and he got paid whenever
somebody got into a fight? That's the business model here.
As Stoller points out in a recent interview with Useful Idiots, the calls
for Silicon Valley to crack down on "misinformation" and "extremism" is
rooted in a basic misunderstanding of how these firms make money. Even as a
cynical or draconian method for clamping down on speech, getting Facebook or
Google to eliminate lists of taboo speakers wouldn't work, because it
wouldn't change the core function of these companies: selling ads through
surveillance-based herding of users into silos of sensational content.
These utility-like firms take in data from everything you do on the Web,
whether you're on their sites or not, and use that information to create a
methodology that allows a vendor to buy the most effective possible ad, in
the cheapest possible location. If Joe Schmo Motors wants to sell you a car,
it can either pay premium prices to advertise in a place like Car and
Driver, or it can go to Facebook and Google, who will match that car
dealership to a list of men aged 55 and up who looked at an ad for a car in
the last week, and target them at some other, cheaper site.
In this system, bogus news "content" has the same role as porn or cat videos
- it's a cheap method of sucking in a predictable group of users and keeping
them engaged long enough to see an ad. The salient issue with conspiracy
theories or content that inspires "societal discord" isn't that they achieve
a political end, it's that they're effective as attention-grabbing devices.
The companies' use of these ad methods undermines factuality and journalism
in multiple ways. One, as Stoller points out, is that the firms are
literally "stealing" from legitimate news organizations. "What Google and
Facebook are doing is they're getting the proprietary subscriber and reader
information from the New York Times and Wall Street Journal, and then
they're advertising to them on other properties."
As he points out, if a company did this through physical means - breaking
into offices, taking subscriber lists, and targeting the names for ads - "We
would all be like, 'Wow! That's outrageous. That's crazy. That's stealing.'"
But it's what they do.
Secondly, the companies' model depends upon keeping attention siloed. If
users are regularly exposed to different points of view, if they develop
healthy habits for weighing fact versus fiction, they will be tougher
targets for engagement.
So the system of push notifications and surveillance-inspired news feeds
stresses feeding users content that's in the middle of the middle of their
historical areas of interest: the more efficient the firms are in delivering
content that aligns with your opinions, the better their chance at keeping
you engaged.
Rope people in, show them ads in spaces that in a vacuum are cheap but which
Facebook or Google can sell at a premium because of the intel they have, and
you can turn anything from QAnon to Pizzagate into cash machines.
After the January 6th riots, Stoller's organization wrote a piece called,
"How To Prevent the Next Social Media-Driven Attack On Democracy-and Avoid a
Big Tech Censorship Regime" that said:
While the world is a better place without Donald Trump's Twitter feed or
Facebook page inciting his followers to violently overturn an election,
keeping him or other arbitrarily chosen malignant actors off these platforms
doesn't change the incentive for Facebook or other social networks to
continue pumping misinformation into users' feeds to continue profiting off
of ads.
In other words, until you deal with the underlying profit model, no amount
of censoring will change a thing. Pallone hinted that he understood this a
little on Thursday, when he asked Zuckerberg if it were true, as the Wall
Street Journal reported last year, that in an analysis done in Germany,
researchers found that "Facebook's own engagement tools were tied to a
significant rise in membership in extremist organizations." But most of the
questions went in the other direction.
"The question isn't whether Alex Jones should have a platform," Stoller
explains. "The question is, should YouTube have recommended Alex Jones 15
billion times through its algorithms so that YouTube could make money
selling ads?"
Below is an excerpted transcript from the Stoller interview at Useful
Idiots, part of which is already up here. When the full video is released,
I'll update and include it.
Stoller is one of the leading experts on tech monopolies. He wrote the Simon
and Schuster book, Goliath: The Hundred Year War Between Monopoly Power and
Democracy, and is a former policy advisor to the Senate Budget Committee.
His writing has appeared in the Washington Post, the New York Times, Fast
Company, Foreign Policy, the Guardian, Vice, The American Conservative, and
the Baffler, among others. Excerpts from his responses to questions from
myself and Katie Halper are below, edited for clarity:
Matt Taibbi: There's a debate going on within the Democratic Party-aligned
activist world about approaches to dealing with problems in the speech
world. Could you summarize?
Matt Stoller: There are two sides. One bunch of people has been saying,
"Hey, these firms are really powerful." This is the anti-monopoly wing.
Google and Facebook, let's break them up, regulate them. They're really
powerful and big, and that's scary. So, without getting in too deep, there's
the Antitrust subcommittee, that's been saying, "Hey, these firms are really
powerful, and they're picking and choosing winners." Usually, they talk
about small businesses, but the issue with speech is the same thing.
Then there's another side, which is, I think, noisier and has more of the
MSNBC/CNN vibe. This is the disinformation/misinformation world. This is the
Russiagate people, the "We don't like that Trump can speak" type of people.
What their argument is, effectively, is that firms haven't sufficiently
curated their platforms to present what they think is a legitimate form of
public debate. They're thinking, "Well, we need to figure out how to get
them to filter differently, and organize discourse differently."
Ideologically, they just accept the dominance of these firms, and they're
just saying, "What's the best way for these firms to organize discourse?"
Taibbi: By conceding the inevitability of these firms, they're making that
concession, but saying they want to direct that power in a direction that
they'd like better.
Stoller: That's right. I mean, there's a lot of different reasons for that.
Some of them are neoliberal. A lot of the law professors are like, "Oh, this
is just the way of technology, and this is more efficient." Therefore, the
question is, "How do you manage these large platforms?" They're just
inevitable.
Then there are people who are actually socialists who think, "Well, the
internet killed newspapers. The internet does all of these things. Also,
there's a bunch of them that never liked commercial press in the first
place. A lot of well-meaning people were like, "We never liked advertising
models to begin with. We think everything should be like the BBC."
So, those are the two groups that accept the inevitability thesis. It's
really deep-rooted in political philosophy. It's not just a simple
disagreement. Then there are people like us who are like, "No, no. Actually,
technology is deployed according to law and regulation, and this specific
regulatory model that we have, the business structures of these firms, the
way they make money from advertising, those are specific policy choices, and
we can make different ones if we want."
Katie Halper: When you say socialist, some may identify as socialists, but
that there's a general group of people who just believe, "We oppose hate
speech and White supremacy," and so we have to make these companies that are
evil, and give them moral authority and a content moderation authority,
which is an inherent contradiction/wishful thinking/inconsistent paradox.
In other words, you're saying leftists, right? Leftist, not liberals, not
neo-liberals, not even liberals, but people who are really would identify as
left.
Stoller: Yes. There's a part of the socialist world that's like, "What we
really want is egalitarianism in the form of a giant HR compliance
department that tells everyone to be tolerant." Right? Then there are most
people who are like, "No. I just don't like wall street and I want people to
be equal and everyone should have a little bit over something," and they
both call themselves socialists.
Taibbi: You and the American Economic Liberties Project have said, there's a
reason why taking Trump off Twitter isn't going to fix the problem, because
you're not fixing those incentives. Can you talk about what those incentives
are, and why they cause the problems?
Stoller: Google and Facebook, they sell ads, right? They collect lots of
information on you and they sell ads and ads are valuable for two reasons.
One, you're looking at them. Two, if they know who you are and they know
information about you, then they can make the ad more valuable. A random ad
space isn't worth very much, if you're showing it to some undefined person.
An ad space you're showing to a 55-year-old man who's thinking of buying a
luxury car, somebody will pay a lot for that ad space, if you know who that
person is and you know that that person has actually been browsing luxury
car websites and reading the Wall Street Journal about how best to liquidate
their portfolio or something to buy a luxury item.
Google and Facebook want to sell that advertising particularly on their
properties, where they get to keep 100% of the profits. If Google sells an
ad on YouTube, they get to keep the money. Facebook sells an ad on Instagram
or Facebook, they get to keep the money. So, their goal is to keep you using
their sites and to collect information on you.
Taibbi: What methods do they use to keep you on the sites?
Stoller: They have all sorts of psychological tricks. Engagement is the way
that they talk about it, but it's like if you go and you look for something
on YouTube, they're going to send you something that's a little bit more
extreme. It's not necessarily just political. It's like if you're a
vegetarian, they'll say, or if you look at stuff that's like, "Here's how to
become a vegetarian," they'll say, "Well, about becoming a vegan?" If you
look at stuff that suggests you're a little bit scared of whether this
vaccine will work, if you search for, "I would want to find safety data on
this vaccine," eventually, they'll move you to like serious anti-vax world.
So, the question that we have to ask is whether you should block crazy
people from saying things, or do something else. Like Alex Jones, for
example, is crazy person or an entertainer, he says things that I don't
particularly like or agree with. The question, though, isn't whether Alex
Jones should have a platform. We've always allowed people to go to the
corner of the street and say whatever they want or to write pamphlets or
whatever.
The question is, should YouTube have recommended Alex Jones 15 billion times
through its algorithms so that YouTube could make money selling ads? That's
a different question than censorship.
Taibbi: Conversely they're not recommending other material that might
discourage you from believing Alex Jones.
Stoller: Right. The other thing is, it's not just that they want to create
more inventory so they can sell ads. It's also the kinds of ads that they're
selling. So, you can sell an ad based on trust. The New York Times or the
Wall Street Journal - I hate using them as examples - they have an audience
and people. They built that audience by investing in content, and then they
sell ads to that audience, and the advertiser knows where that advertising
is going and it's based on trust.
The alternative model which we have now is simply based on clickbait. It's
just, "Generate as many impressions as possible, and then sell those
impressions wherever the person is on the web." That creates a kind of
journalism, which is designed to get clicks or not even journalism. It's
just you're creating content just to get engagement and not actually to
build trust.
So, what this business model does, we call it surveillance advertising, but
it's an infrastructure player, a communications player manipulating you so
that they could put content, engage content in front of you. What that does
is it incentivizes a low trust form of content production. It both kills
trusted content producers, a.k.a. local newspapers, because you no longer
need to be able to put advertising in the Pittsburgh Post-Gazette, or
whatever. You can just geotarget those people through Google and Facebook.
You can get some Eastern European to falsify stories and people will click
on that.
So, it kills legitimate newspapers and it creates an incentive for low trust
content, fraudulent content, defamatory content, whatever it is that will
keep people engaged and is often fraudulent. It hits really local newspapers
and niche newspapers the most, so Black-owned newspapers and also newspapers
having to do with hobbies. The actual issue is more about niche audiences
themselves, and the kind of low-trust content that we're encouraging with
our policy framework, versus what we used to do, which we would encourage
higher trust forms of content.
Taibbi: How would you fix this problem, from a regulatory perspective?
Stoller: The House Antitrust Subcommittee had a report where they
recommended what we call regulated competition. That would say, "Okay. You
break up these platforms in ways that wouldn't interfere with the efficiency
of that particular network system." So, Google and YouTube don't need to be
in the same company, you could separate them out.
There are ways that you'd have to handle the search engine. You couldn't
split Facebook into five Facebooks, because then you wouldn't necessarily be
able to talk to your friends and family, but you could separate Instagram
and Facebook easily. You could force interoperability and then split up
Facebook if you want to do that. So, you could separate those things out and
then ban surveillance advertising for a starter.
Taibbi: What would that do to content if you ban surveillance advertising?
ANd how would that work?
Stoller: It would force advertisers to go back to advertising to audiences.
So, they would no longer be able to track you as an individual and say, "We
know this is what you're interested in." They would go to what's called
contextual advertising, and they would say, "Okay. If you're on a site that
has to do with tennis, then we'll advertise tennis rackets on that site
because we assume that that people are interested in tennis rackets."
That's called contextual advertising, versus the current system: you read an
article about tennis in a tennis magazine and the platforms say, "Oh, that's
expensive to buy an ad there, so we'll track you around the web and when
you're on Candy Crush, we'll show you a tennis racket ad." That's the
surveillance advertising model we have. That pulls all the power to Google
and Facebook who are doing all the tracking, versus the contextual ad where
the power is actually with the tennis racket site that has the relationship
with the people interested in tennis.
Taibbi: So, the idea would be you would create a sort of a firewall between
the utilitarian functions of a site like Facebook or Google, that provide a
service where either you're searching for something or you're communicating
with somebody, and they wouldn't be allowed to take that data from that
utility-like function to sell you an ad?
Stoller: That's right. Germany is hearing a court case saying that you can't
combine advertising from Facebook, Instagram, and WhatsApp, and then third
parties to create a super profile of someone and show them ads. They were
saying that's an antitrust violation. There's a court hearing on that, but,
more broadly, that's what you have to do.
Ultimately, what we would want is we would want to have subscription-based
communication networks paying for services. This is something that's worked
for thousands of years. I give you something in value, you give me money.
It's an honest way of doing business. If I don't value it enough to give you
money, then I won't get it.
If people are like, "Oh, I don't want to pay for Facebook, or I don't want
to pay for YouTube," or whatever it is, that makes no sense. You're already
paying. You're either paying with a Friday night you spend surfing YouTube,
where they sell a bunch of ads and you give up your Friday night - or you
pay with money, and it's an honest transaction, and it's in the long run, a
lot cheaper and more honest method of payment.
For more of this interview, check out UsefulIdiots.Substack.Com in the next
days. For Stoller's writings on the subject, see here.
This post is only for paying subscribers of TK News by Matt Taibbi.
Like & Comment
Give a gift subscription
Facebook
Twitter
Instagram
Youtube
C 2021 Matt Taibbi Unsubscribe
548 Market Street PMB 72296, San Francisco, CA 94104