What Is Antitrust Really Against ④: The Fear of Big Companies Controlling Information and Speech Is Coming True

Whether you pay attention to a crowd of people talking nonsense is your business. Whether you can see what they say at all is not.

In a world sliding into two opposed camps, consensus gets harder and harder to reach.

How much of a market a big company must control before it counts as a monopoly is a question lawmakers answer differently by region and by politics. Google has already collected three rounds of European antitrust fines totaling 8.2 billion euros, while several investigations into the same competitive tactics in the United States ended in findings of no wrongdoing. In China, being referee and player at once has become so ordinary that it barely gets mentioned.

But on the other flashpoint of antitrust — the fear that tech companies control speech and the spread of information — America's two endlessly quarreling parties turn out to agree.

On March 8, Senator and Democratic presidential candidate Elizabeth Warren announced an antitrust plan to break up the big tech companies, and Facebook promptly took her break-up ads down. Ted Cruz, a Republican leader who ran for president in 2016, quickly voiced support, saying tech companies "have too much power to silence speech." A year earlier, Cruz had already tied tech antitrust to the First Amendment: "Everyone should have freedom of speech — that is a right the First Amendment gives Americans. Tech companies may have political leanings, but they cannot decide other people's freedom of speech. We have an obligation to protect that right with antitrust or other legal tools."

More recently, Josh Hawley, a rising Republican star and the youngest senator in American history, has begun calling for limits on big tech's control of speech on its platforms. "The First Amendment drew the boundary on censorship a century ago. Tech companies should not get to censor and bear no regulation for it."

The fear that big companies will use a monopoly position to control speech and the flow of information goes back to the birth of antitrust itself. In 1900, the Illinois Supreme Court ruled that the Associated Press had to supply news to every organization, after the AP cut off customers that ran stories from rival wire services. The court held that the AP, in controlling the news, amounted to a public platform and was using its dominant position to restrict the free circulation of news.

A century later, the focus has moved to Silicon Valley. The two most dazzling stars of internet entrepreneurship, Google and Facebook, control what billions of people read, watch and share. The power over information that comes with monopoly scale is now itself a reason for antitrust.

Starting From "Don't Be Evil": How Big Companies Accumulated Control Over Information

On September 5, 2018, an extra empty chair sat at the witness table on Capitol Hill. The nameplate in front of it read Google. To its left and right sat Facebook's chief operating officer and Twitter's founder. The tech representatives took three hours of questions from the Senate Intelligence Committee. Only Google was missing.

The standard-bearer of "Don't be evil" had turned into an information controller that worries the world's only superpower.

For a long stretch after it was founded in 1998, Google was the American dream by the template. Two shy doctoral students, one of them newly arrived in the United States from a communist country that had just fallen, with none of the black marks of the previous generation of tech stars — refusing to acknowledge a daughter, or grabbing a partner's shares after his terminal diagnosis.

Google's free corporate culture drew in countless young engineers: pets allowed at work, free perks, encouragement to invent and to develop as individuals, and the fortunes that came with a rising stock.

More important, Google held to "Don't be evil." Larry Page and Sergey Brin, the two founders, resisted advertising at first. In his Stanford paper, Page wrote flatly that "if you sell ads, a search engine goes bad." Even when growth demanded more ad revenue, Google wanted to keep something intact.

"Don't be evil" stood for exactly that kind of holding the line. To Google's early employees, other search companies were cheating users by selling their search results. Google would not make its money from the results; it put ads to the right of them and labeled them as clearly as it could.

Google meant to index the world's information and to serve up accurate, useful results without letting profit bend them. By the 2004 listing, the line had made it into the prospectus, and the letter to Google's investors said: "Don't be evil. Even if it means giving up short-term gains, a company that does good for the world will, in the long run, be better rewarded."

But Google's grip on information kept tightening.

In 2006, after Gmail and Maps, Google bought the one-year-old YouTube for $1.65 billion. Two years later, Chrome arrived and began eating Microsoft Internet Explorer's share. At almost the same moment, the HTC G1, the first phone running Android, went on sale.

Google's business model has always been to offer better internet services, collect more information about the people using them, and aim ads more precisely with what it learns.

Tim Cook likes to run Android down by saying Google sells user data. Google never sells data outright. It simply knows all of it, and puts advertisers' messages in front of the right people.

That is how ads reach each person. It is also how information does.

Google, now the largest artificial-intelligence company in the world, began testing personalized search in 2007: algorithms read a user's search behavior and returned results better fitted to them.

The trouble started once algorithms ran the results. In every presidential election since 2008, the political tilt of Google's search results has been disputed.

Because Android and Chrome hold monopoly-sized shares, and because Google's search is genuinely good, Google decides in practice what appears in front of people when they look something up.

Eyetools, a firm that studies how eyes move across a page, went further: the higher a result ranks, the more attention it gets. The top three Google results draw 100% of readers' attention; the results at the bottom get 20%.

Algorithms decide what information users get. And with Google's purchase of YouTube and the launch of Google News, still more of that information belongs to Google.

Facebook grew up much the same way. In a public speech in 2009, Mark Zuckerberg — 25 years old and five years into running a company — urged people to put their lives online and promised a public social network linking everyone on earth. Facebook now has 2.3 billion monthly active users, and the average person spends more than 50 minutes a day on it — not counting the other products it has bought.

Facebook added algorithms early too. The News Feed launched in 2006. News, friends' posts and cat-and-dog pictures were ranked by algorithm, pulling users back to refresh again. Because a social network is consumed passively, its algorithm decides what each person sees even more directly than search does.

In Antitrust History, Two Fears Drive Everything: Control Over Speech and Restraint of Competition

We have already covered, earlier in this series, the fear that whoever holds the infrastructure will throttle competition.

Besides protecting competition and keeping a few companies from setting prices, guarding against big companies' control over speech and information is the other thread running through antitrust history.

In the late 19th century the United States ended slavery and industrialized fast. Monopolies formed across daily life — railroads, farming, tobacco, grain, oil — and the news business was no exception.

In 1892, the Associated Press of Illinois, the AP's predecessor, was founded. Its president, M. E. Stone, paid heavily for exclusive news-exchange contracts with Europe's three big wire agencies and cornered the supply of foreign news.

Control of foreign news let the Associated Press of Illinois beat its largest rival, United Press, and let it demand that partner papers run its copy and no one else's.

When Chicago's Inter Ocean ran another agency's stories, Stone cut it off. The paper went to court, and after two years of proceedings the Illinois court ruled, in 1900, that the wire service had to supply copy to any customer without discrimination.

The reasoning was that the state supreme court considered press freedom protected by the First Amendment, which made the AP something like a public medium, obliged to supply everyone. Unwilling to live with the ruling, Stone dissolved the Illinois service and reorganized it in New York to get around the judgment. Antitrust law was still full of holes.

Antitrust rules watched the dangers of media monopoly from the start. In 1914 the Clayton Act and the Federal Trade Commission Act arrived, filling out the first antitrust statute, the Sherman Act.

The Clayton Act flatly banned share acquisitions made "to weaken a competitor or create a monopoly." That stopped the big newspaper groups from swallowing small ones at will. In 1893, the biggest of them, the Scripps chain, was buying an average of six local papers a year, and controlled the newspaper content of 30% of America's industrial cities.

News monopolies got that much attention because free speech carried by the press was treated as the foundation of democracy. Thomas Jefferson, the main author of the Declaration of Independence and the third U.S. president, once said, "Were it left to me to decide whether we should have a government without newspapers, or newspapers without a government, I should not hesitate a moment to prefer the latter." Hugo Black, the Supreme Court justice appointed by Franklin Roosevelt, held that the chief function of a free press was to keep the government from deceiving the people.

Newspapers, radio, telephone, telegraph, then television — the medium kept changing, but antitrust's guard over the free flow of information held into the 1990s.

In 1934, the U.S. government passed the Communications Act, and the Federal Communications Commission followed, setting the terms of telecom regulation for the next half-century.

From 1940 on, the FCC, to stop television stations from cannibalizing and buying one another, issued a series of strict rules to hold the industry in check. Antitrust rules held that one broadcaster could own no more than 12 stations nationwide, could not own two stations in any of the country's 50 largest cities, and could not own a radio station and a cable system at the same time.

The cable market opened up in the 1980s, and CNN, Discovery, ESPN and HBO were all born there. To keep cable operators from deciding what people could see, the Cable Television Consumer Protection and Competition Act wrote in "must-carry": local stations that covered local life and local politics had to be given airtime on cable systems.

Just as Rockefeller used a railroad monopoly to take the oil market, a cable operator that controlled the channel could decide what was broadcast and squeeze local stations out. Programs those stations made drew no ratings, leaving them no choice but to sign exclusive deals with the operator.

The purpose of the rules was plain: the pipe that carries information must not control what travels through it, so that every political opinion can get out.

Mergers between companies that held both the information and the pipe were policed hardest. For a long time, cable operators that owned the distribution networks could not buy television stations, and film studios could not own theaters.

For newspapers, the Newspaper Preservation Act of 1970 let two companies share advertising revenue to stay alive, on condition that they keep separate editorial staffs, so the range of news and information would not narrow.

Tim Wu, the Columbia law professor who put forward the theory of "net neutrality," writes in The Master Switch: The Rise and Fall of Information Empires that for decades the U.S. government watched the flow of information closely, determined that no single company should come to own information and speech.

That changed in the 1980s. The Reagan administration pushed economic freedom, loosened monopoly enforcement and shrank the government's role in the economy. In 1986, Daniel Oliver, the chairman of the Federal Trade Commission, gave a speech to the American Bar Association that showed the new tone: "The enemy of public policy is excessive government intervention; the damage antitrust enforcement does to business activity should be limited."

After that, breakup litigation against big companies flared only once more, against Microsoft under Clinton. George W. Bush's administration was more forgiving of big-company monopolies still. By 2008 Obama was elected, and was called the first internet-celebrity president; from the start of the campaign he used Facebook and other social networks to gather votes.

The Power Over Speech Is Not Always Wielded by the Tech Companies Themselves, Which Makes Antitrust Harder

The flow of information across the internet has gradually ended up in the hands of two companies. Netcraft.com, after analyzing traffic to 1.4 billion websites, found that Google and Facebook, together with YouTube and Instagram, which they bought, directly steer more than 70% of the world's internet traffic.

Traffic to news sites shows that control more plainly still. Since 2013, referrals from Google and Facebook have accounted for more than half of all traffic to media sites. China is the exception: companies there hold shares just as dominant, but they are not the ones controlling information.

Every four years, Google's and Facebook's political leanings become an argument. By 2016 the public doubts had grown louder.

In 2016 the independent outlet SourceFed claimed Google's search engine carried a "political bias" and had manipulated results to bury negative information about Hillary Clinton, the Democratic presidential candidate. Google answered that the algorithm had decided on its own.

Early last year the Cambridge Analytica leak broke. Data on 50 million Facebook users had been used by the data firm Cambridge Analytica, and was thought to have helped Trump canvass votes during the presidential election. Taking social-network data and targeting voters with it had been done in Obama's day too; the difference is that Obama had Facebook's consent — Facebook co-founder Chris Hughes joined the Obama campaign outright.

After Trump won, Reuters reported that Google employees had discussed internally changing the algorithm to suppress conservative news outlets. Sundar Pichai, Google's chief executive, testified before Congress seven times last year, repeating each time that Google is politically neutral.

Facebook drew the same charge. Former employees described how the company manually shaped opinion during the 2016 election: left-leaning staff filtered out conservative stories. The Senate Commerce Committee promptly wrote to Zuckerberg demanding an account of how Facebook's Trending Topics section worked.

What is different from the fear of a century ago is that it is hard to picture Mark Zuckerberg or Larry Page issuing a secret order for a political hit. But tech companies are not the only ones using recommendation algorithms to shape what people see.

Foreign interference in the American election landed Google and Facebook in far deeper trouble.

Zuckerberg first denied that Russian agencies had any part in the Cambridge Analytica leak. Then the GRU's plan to interfere in the election came to light, and he changed his answer. Across two days and ten hours of questioning in Congress, Zuckerberg admitted Facebook had failed to spot the ads Russia bought in time, and that fake news, foreign election interference and hate speech followed.

Fake news is the commonest way to steer speech on social networks. The New York Times interviewed people who make YouTube videos with fake-news headlines. They live off the recommendation algorithm and the clicks it tricks out of viewers; the ones interviewed could make $200,000 a year.

Not every fake-news producer has a political aim. Some simply know readers like a lurid story, and that lurid stories sell more ads. Inside an information flow controlled by social networks and search engines, true and false grow harder to tell apart — from elections and Brexit in rich countries to ethnic conflict and revenge killings in remote ones.

Plenty do have political aims. Working the same Google algorithm, Reddit users linked the word "idiot" to photographs of Trump in bulk, until a search for "idiot" returned nothing but Trump — which it still does. ISIS even ran a new-media department, opening accounts on Facebook and Twitter by the thousand to preach jihad and upload beheading videos.

There is no evidence that Facebook and Google control the internet's information maliciously. The facts do show that two companies serving billions of people a day have enough influence to do harm.

The fake-news scandals put tech companies under enormous pressure, and under it they began exercising the power to censor.

On March 15, a gunman killed 50 people in Christchurch, New Zealand. He livestreamed the killings on Facebook. Facebook's artificial-intelligence video filters failed to catch the broadcast, and afterward neither Facebook nor Google could stop copies of the video from resurfacing.

Accused of policing too little, Facebook policed harder, blocking even the clips used in news reports.

The same sequence has played out in one Facebook and YouTube controversy after another. Since 2017 the number of Facebook content moderators has tripled, from 10,000 to 30,000 — about as many as the company's full-time employees worldwide.

Where the line falls between what belongs on the internet and what should be deleted is hard to draw. The more serious problem is that the judgment still rests with algorithms.

In the most recent cases — far-right speech, and the anti-vaccine movement — Amazon, Facebook and YouTube purged the material one after another. That has become a new kind of censorship.

The modern anti-vaccine argument began with a paper in The Lancet, the famous medical journal: the physician Andrew Wakefield's study concluded that vaccination gave children colitis and might even bring on autism. The argument caught on, and its supporters believed medical companies were hiding what vaccines did in order to make money.

As the fight sharpened, tech companies could not stay out of it. Amazon banned anti-vaccine documentaries outright; Pinterest and Facebook deleted anti-vaccine material across their platforms. YouTube did not delete the videos, but made it less likely they would be recommended.

The anti-vaccine movement pushed companies that profess to believe in free speech into blanket bans; some of the biggest handled it almost the way they handle Nazi content or child pornography.

That worried the American Medical Association (AMA), which wrote to five tech companies arguing that the giants should do their part to make sure people can find scientifically sound information about vaccination, rather than deleting content across the board.

Tech companies are getting used to managing how information moves.

Facebook tried rating news outlets itself to deal with fake news. A new doubt followed: why should Facebook be the one to decide which outlets' information deserves to travel? Alexis C., a deputy editor at The Atlantic, wrote in criticism that "the ratings get in the way of press freedom."

And a leaked internal Google document shows that Google, too, is coming to accept the role of censor.

Jason Pontin, the former editor-in-chief of MIT Technology Review, was asked by Google to answer one question: how do you screen out harmful content while keeping conversation free?

The 86-page internal slide deck worked through a large number of journalists' articles, along with academic and critical interviews and source material, and reached a conclusion: under ever harsher criticism, Google, Facebook and Twitter can only censor and manage more tightly, and can no longer sustain the free-speech ideal the internet was founded on.

Bit by bit, the big tech companies that once believed in "Don't be evil" and in the free-speech principles of the early internet are settling into their new work as censors, serving whichever master the situation calls for.

Cover image source: Wiki