What Is Antitrust Really Against ⑤: Google Is Back Before Congress Over Its Control of Information — What Charges Have Piled Up Against It?

The problems are being laid out one by one; a workable fix has yet to appear

Executives from America's internet giants are on their way back to Congress for another round of questioning.

On April 10, spokespeople for Google and Facebook will take questions at a hearing of the U.S. House Judiciary Committee and explain why video of a white-supremacist gunman's killing spree in New Zealand spread so widely on their platforms. A day later, they face the U.S. Senate, on a subject stated plainly — "Stifling Free Speech: Technological Censorship and the Public Discourse."

The two subjects together show the bind tech companies are in, against the backdrop of the antitrust debate, because of the control they hold over the flow of information.

In the Democrat-controlled House, Judiciary Committee Chairman Jerrold Nadler accused the tech giants of tolerating the spread of hate speech, saying "tech companies have plainly become a conduit for hate speech." Kristen Clarke, president of the Lawyers' Committee for Civil Rights Under Law, submitted a report to Congress arguing that social networks are still a key place where many violent white-supremacist groups organize and recruit.

In the Senate, the charge runs the other way: too much censorship, free expression stifled. Republican Ted Cruz, who chairs the Senate hearing, said "the new hearing seeks to discuss the widespread bias in speech and content censorship at tech companies, a fear held by many Americans." A year earlier, Cruz raised the same point, warning that "tech companies cannot override the freedom of speech guaranteed by the First Amendment."

Unlike the antitrust fines the EU has handed out one after another, the big problem Google and Facebook face at home is not that they restrict commercial competition or squeeze out small companies. It is the enormous influence they hold over speech and the spread of information.

Bad news about Facebook has never stopped, and people have long since gotten used to it. Google's position has changed enormously over the past year and a bit. Since its first questioning in August 2018, Google has been called before Congress four times. Less than four months after CEO Sundar Pichai testified in person late last year, a new and more serious round of debate is already under way.

Criticism from outside keeps coming, and Google has had trouble inside as well. An ethics board set up to handle the moral questions raised by artificial intelligence lasted a week: after a wave of employee protest over one controversial member, the company hastily dissolved it. Mass employee protest shutting down a project is not new at Google — it has already happened once before in the past six months.

Few companies reflect the growth and the controversies of the tech industry over the past 20 years better than Google.

Google came out of a textbook American Silicon Valley dream. Six years took it from a garage to Nasdaq. The year its two young founders took the company public under the creed "don't be evil," Mark Zuckerberg was 20. A whole generation of tech companies has taken it as the model — office culture, equity structures, OKR reviews.

Worries about Google's monopoly over information go back a long way. In 2005, a New York district court halted Google's settlement with book publishers. The presiding judge, Denny Chin, held that Google's ability to control information about books online would entrench its monopoly in the online search market.

When Margrethe Vestager, the EU's antitrust commissioner, had just taken office in 2014, she said: "The amount of data Google controls will raise a whole set of social challenges, far beyond economic ones." But at the time, those in power cared more about the economics, and Vestager went on to help hand down three fines totaling 8.2 billion euros.

Now Congress's more serious, more frequent accusations against Google come from worry about the control of speech, not about market competition.

In December 2018, Google CEO Sundar Pichai stood before a congressional hearing for the first time. The first question thrown at him by House Republican Leader Kevin McCarthy was: "Are American tech companies a tool of freedom, or a tool of control?"

In 15 years, worry about Google has climbed from search-engine competition all the way to thought control. That climb is one measure of how much society, and tech companies, have changed.

In 2005, the most popular phone was still a Nokia, and the iPhone was two years off. The internet was, in effect, a thing that existed only once you lit up a computer screen. People came to this newly settled territory of some 200 million so-called netizens in order to get information.

In 2019, the iPhone and Android are so widespread and so refined that phone makers no longer know how to talk anyone into buying more phones. Across most of the planet, more than 4 billion people are online in real time through their phones, and five hours a day is nothing remarkable.

The reach of tech companies' products has grown with them. Google Search decides what 2 billion people can see. Android's user base also passes 2 billion. Facebook has 2.3 billion users, and that one app alone takes more than 60 minutes of the average user's day. Voices from all over the world have been funneled onto a handful of tech platforms.

Once the internet went from a community that very few people took part in to something that seeps into everyone's daily life, it could no longer be the ideal world its pioneers wanted. Everything good and bad in the real world is now mapped onto it, and sped up. But what shapes all of this, day after day, is no longer government. It is a comparatively small number of tech companies that have not been around long. As Zuckerberg's latest open letter says, this is also a job that gets harder to do well.

Faced with tech giants whose reach is already frightening and still growing, regulators, politicians, activists and even the companies' own employees keep drawing up charge sheets. And the remedies they call for have escalated from fines and promises to strict regulation, or a breakup.

Using Google as the example, we have gone through the charges the outside world has laid — from 2005 to today — against this tech giant and the industry it stands for.

Some of these positions contradict one another. And more than once, an attempt to fix one problem has bred a new one.

Google's first antitrust trouble began with book copyright — and with the internet company's usual habit of acting first and asking later

Google's secret plan to scan every book in the world began in 2002, under the internal code name "Moonshot." A company that had only just started making money spent $50 million a year on a project with nothing to do with its own business. At the time, scanning the University of Michigan's library was thought to need 1,000 years; Google believed its own algorithm could finish in six, and would keep getting faster.

The trouble lay outside the technology. The authors and publishers who held the copyrights quickly saw that Google had ransacked the libraries without any permission at all. On September 23, 2005, with only a few million books scanned, Google was sued for the first time by the Authors Guild and the Association of American Publishers — but Google talked the authors' and publishers' groups around with the market its technology opened up: "Out-of-print books gathering dust on library shelves are worth nothing; open search can bring in new revenue."

The protests against the settlement were fierce. Robert Darnton, then director of the Harvard library, asked: "Do we really want to hand over our greatest libraries to a giant corporation that can charge us whatever it likes?"

The Justice Department's antitrust division stepped in. The question in the case shifted from whether Google had the right to scan books to whether it was building and monopolizing an online book market.

The department had a further reason to oppose the settlement. Its closing report argued that Google's method — deliberately infringing copyright first, then sorting things out through litigation — should not be encouraged. The courts rejected the settlement three times. In the end the case took eight years: in November 2013, a New York state court ruled that Google's book scanning was not infringement, closing it out.

Google Books existed because Google was chasing its most basic goal, "to index all information and let everyone reach it for free." Along the way, it did not wait for copyright holders to grant permission or push things forward. It acted first and negotiated terms once it had scale.

That approach runs through nearly every case of internet companies pushing into real-world problems — Uber and Didi turning ordinary car owners into full-time drivers, Facebook's privacy boundaries, food delivery with no formal employment relationship at the start, video sites keeping their guard down on infringement to hold startup costs low. In the phrase Facebook later put on its office wall: Move Fast and Break Things.

Similar problems turned up later in Google's Street View privacy scandal and its advertising scandal.

On May 9, 2010, Hamburg's data protection authority in Germany found that Google's Street View cars had been collecting Wi-Fi data from homes along their routes. Called in to explain, Google insisted the cars collected only Wi-Fi SSIDs and MAC addresses and stored no other sensitive data. Then-CEO Eric Schmidt called it the private act of one employee: an engineer had slipped a piece of code into the Street View software, in "clear violation" of procedure. As the investigation went deeper, Google had to admit it had stored sensitive information after all.

On October 25, 2010, Google acknowledged on its official blog that for three years its Street View cars had been collecting information users sent over unencrypted Wi-Fi networks along the route — email account passwords, text messages, phone numbers, medical records and other private data, 600 gigabytes of it. The U.S. FTC's investigation report also found that Google's executives and engineers had both reviewed the code in question and had not fixed the problem in it.

In 2009, David Whitaker, an American businessman selling fake drugs online, put steroid labels on bottles of purified water, advertised them through Google AdWords, and sold them from Mexico to U.S. customers at $1,000 a bottle.

After Whitaker was arrested, he agreed, in exchange for a lighter sentence, to do another deal with Google under Justice Department surveillance. During the sting, Google's ad sales staff knew he had no license to sell drugs and went so far as to offer to design him a website to help promote the fake pharmaceuticals. Google was ultimately fined $500 million.

After each of these episodes Google took a heavy fine, then built a mechanism to fix the problem and win back forgiveness. It became standard practice across the internet industry.

But as their scale and their grip have grown, the new problems the tech giants face are no longer so easy to settle.

Then it widened to tech giants using monopoly power to hold off competition — the EU has already issued billions of euros in fines for it

In February 2010, the EU concluded that Google was deliberately pushing down rival price-comparison services in its search results. Eight months later, it opened a formal antitrust investigation. Google had changed the way search results were displayed, so that a search for a product pointed straight to product pages rather than to the many price-comparison sites then popular in Europe.

The European Commission analyzed the results pages of 1.7 billion Google queries and, after seven years of investigation, concluded that Google had suppressed competitors. In June 2017, Google was fined 2.7 billion euros and promised to change how it handled search.

A similar charge applies to restaurant search: Google would rather send results to the information page and booking service it built for restaurants itself than to pages run by its rival Yelp.

Google's more obvious monopolistic behavior is bundling services through Android.

On April 9, 2013, the EU brought an antitrust case over the Android operating system, arguing that Google used Android to bundle and preinstall its own apps — browser, photos, maps, cloud storage. The European Commission confirmed that Google had indeed used exclusivity agreements and subsidies to monopolize preinstallation among manufacturers and carriers. In July 2018 the EU fined Google 4.3 billion euros and required it to offer the app store and the operating system separately.

Barry Lynn, author of Cornered: The New Monopoly Capitalism and the Economics of Destruction, argues that Google's behavior is much like a railroad steering passengers into its own shops: control the basic service, and you can hold off your competitors.

But each of these rulings has its critics. Forced bundling of Android apps is a fairly clear case. Yet when someone searches for a product or a shop, why shouldn't the results show the information directly instead of pointing to a comparison site or a review site? As the independent analyst Ben Thompson has put it: "When a consumer searches for a pair of boots, why can't the search engine just show the cheapest link, instead of a row of price-comparison services?"

The U.S. FTC investigated and reached the opposite conclusion from the EU's: Google's monopolistic conduct had not harmed consumers. But the EU's approach has supporters in the United States.

In 2017, Democratic Senator Al Franken called for applying to tech companies the same neutrality principle the government requires of carriers. In a column for The Guardian he wrote: "Just as carriers do, Google, Facebook and Amazon should always stay neutral in how they handle lawful information and business models on their platforms."

Elizabeth Warren, a Democratic candidate in the 2020 presidential race, describes the tech companies that hold the internet's entrances and platforms as the railroads of their era, controlling the arteries of transport — able to move a competitor's logistics costs and take hold of the market that way. She has proposed splitting tech companies into infrastructure services and products, so that a tech company cannot be referee and player at once.

But what really ran Google into antitrust trouble in the United States — a country now capable of deciding the limits of a tech company's business, or of breaking one up — is the control of speech.

Personalized algorithms decide what you see on internet platforms — and they stand accused of letting extremism and child pornography spread

In 2007, Google began testing personalized recommendation: algorithms read a user's searches and video preferences and suggested results that suited them better.

The scale grew fast. Eight years ago, Google passed a billion visitors for the first time. Today it holds 86% of the world's search market — and that is with China left out. A garage startup in Silicon Valley became a giant that shapes what more than 2 billion people see. YouTube, which Google bought in 2006, has grown into a platform of more than 2 billion users, each watching more than an hour of video a day on average. Recommendation turned people from active searchers into passive consumers, pulled along by one suggested item after another, refreshing without end.

There was no stopping what came with it: false, extreme, pornographic and bloody content began appearing on these platforms in bulk. Around moderation, Google became the focus of the argument over the control of speech.

Not many people question whether child pornography and ISIS execution videos should come off a platform. In 2014, the war correspondent James Foley was killed in Syria, and the video of his beheading was uploaded to YouTube and quickly taken down.

Dan Gillmor, editor of The Atlantic, wrote that YouTube had done the right thing, but he also asked a question: removing a beheading video looks straightforward, but not everything is that black and white, so where is the boundary of speech control, and should large corporations be the ones to draw it?

What is clear is that Google's algorithm does not set out to spread extremist video. But personalized recommendation naturally hands traffic to more of the same.

Jack Nicas, a reporter at The Wall Street Journal, ran a test: when a user's choices show a political leaning, YouTube tends to recommend videos that answer to that leaning, and often ones with more extreme views.

Nicas points out why the algorithm works that way. Tech companies need stimulating content to keep users clicking, which is how a platform collects large amounts of their time.

Algorithms cannot solve every problem in moderating malicious content. In the recent New Zealand killings, YouTube said it would remove all video related to the shooting, yet footage kept circulating on the platform. Facebook was the platform the killer livestreamed on, and the broadcast was only shut down after police told the company about it.

Last year, when Google CEO Pichai was questioned on Capitol Hill, Jamie Raskin, a Democratic member of the antitrust subcommittee, argued that when Google's communication tools are used to promote violence, it becomes a serious public problem. Pichai then conceded that Google still has more work to do in writing and enforcing its policies on hate speech and other offensive content.

Social platforms now take hate speech down faster than they used to. In 2018, Facebook, Google and Twitter removed 72% of illegal hate speech, reviewing 89% of it within 24 hours, against 40% in 2016. But at the scale the internet giants reached in 2018, the remaining 11% is still a huge number. Now, with video of the Australian gunman's attack spreading, a new round of questioning begins this week.

Censoring extremist content then ran into the charge of censoring too much

In 2008, Joseph Lieberman, chairman of the Homeland Security Committee and the running mate who lost with Al Gore to George W. Bush, wrote to Google demanding it remove hundreds of videos uploaded to YouTube by al-Qaeda supporters, among them weapons-making tutorials and videos inciting war.

Lieberman told Google to do the right thing without hesitating: "Removing Islamist terrorist videos is a straightforward task." In the end Google did not comply. It took down 80 of the videos and left most of the rest up.

The ACLU (American Civil Liberties Union), founded in 1920 and speaking out on racial discrimination since the days of American segregation, protested against Lieberman. Censoring terrorist content on American soil, the ACLU argued, could be the opening act of controlling speech on the internet.

The conflict broke open in 2017. On August 12 that year, far-right demonstrators clashed with counter-protesters and blood was shed. Soon after, The Daily Stormer, a website promoting Nazism, had its domain registration revoked by GoDaddy and Google, the world's two largest domain registrars. It was the first website mainstream Silicon Valley companies shut out entirely without any clear illegal act.

As social tensions sharpened, people grew less and less willing to tolerate speech they disagreed with. The neutral stance Silicon Valley had always held began, quietly, to shift. What got taken down widened from beheading videos and pornography to content carrying different values.

The ACLU and the civil rights group EFF (the Electronic Frontier Foundation, which stresses the importance of civil liberties and holds that government should not have too much power) both published posts condemning tech companies for collectively silencing neo-Nazi individuals and websites. The methods used to block hate speech, they argued, would in the end be turned against the civil rights groups fighting for freedom. You can read our earlier report on the episode here.

If Nazism is a relatively black-and-white question, this year's fight over anti-vaccine censorship is not.

The anti-vaccine argument has run for more than 20 years. Its supporters hold that vaccination causes autism in children, and that drug companies have hidden this for profit and funded the studies claiming vaccines do no harm. Its opponents say there is no medical evidence for any of it, and that the claim will only raise child mortality. As the two sides grew louder, the tech companies were finally pulled in. The more extreme responses — Pinterest, Facebook, Amazon — blocked anti-vaccine content outright. YouTube did not remove the videos but added warnings to disputed content.

That worried the American Academy of Pediatrics (AMA), which wrote to five tech companies saying the giants should do their part to make sure consumers can get sound scientific information about vaccination, rather than deleting content with a blunt instrument.

After the 2016 election, liberals blamed tech platforms for spreading fake news that interfered with the vote

In November 2016, a Google search for the final election result pointed to a fake-news link claiming Trump had beaten Hillary Clinton in the popular vote. Google later admitted it had failed to keep the fake news out.

The New York Times interviewed the man who made that piece of fake news. His reason: hot political news brings higher page views.

Fake news was first seen as a way to grab attention and turn page views into ad money. But as tech companies' services grew more powerful, fake news and fraudulent ads were found to have been used to interfere in the 2016 election.

In September 2017, Google found that Russia's GRU (Main Intelligence Directorate) had spent tens of thousands of dollars during the earlier presidential election buying ads to put out disinformation on YouTube, Google Search, Gmail and other Google products.

Within a month, Google and Facebook were called to three hearings, on whether they had knowingly accepted political ads placed by foreign powers (illegal in the United States) and whether they were capable of stopping fake news from spreading.

The first to level the charge were, naturally, the liberals furious at Trump's election. Sheldon Whitehouse, a Democratic senator and former Rhode Island attorney general, argued that tech companies had not tried hard enough on fake news, and had helped put Trump in office.

In March 2018, Google announced a $300 million investment to deal with fake news, planning algorithm updates that would surface more authoritative content; Facebook said it would rate media outlets and show more content from the big ones.

That landed Google and Facebook in a fresh round of doubt: why should tech companies decide the ranking of one news outlet against another? The Guardian noted in an editorial that giving more weight to the large news sites would shrink the reach of small publishers further still.

At the same time, conservatives believe Google is actively suppressing their speech

In the end, Democratic liberals and the Republican conservatives Trump represents both accuse tech companies of bias in what they carry. After Google and Facebook appeared at the fake-news hearings, the American conservative magazine National Review commented that the attitude of liberal media and Democrats toward Silicon Valley depends on whether Silicon Valley can keep them in power: as long as it helps the Democrats and the liberals, whatever trouble Silicon Valley runs into will draw no blame.

And Trump, thought by many to have gained more from fake news, accuses the tech giants of spreading bias too.

In December 2018, Republican Congresswoman Zoe Lofgren put a question to Google CEO Pichai, sitting in front of her: "Why does searching the word 'idiot' bring up a photo of President Trump?"

Pichai maintained that the algorithm produced it, but lawmakers' suspicion that Google suppresses conservatives did not die down.

Lamar Smith, a conservative Republican congressman from Texas, argued that there is undeniable evidence Google suppresses conservative speech. His examples: the front page of Google's results holds not one link to a conservative site, and content about immigration law gets flagged as hate speech.

A study by the online search-marketing firm CanIRank.com also found that among the 50 most recent searches for political terms on Google, the pages shown leaned liberal more often than conservative.

In 2016, the independent pop-culture news site SourceFed released a video giving examples that suggested the autocomplete in Google's search service favored Hillary Clinton, the Democratic presidential candidate.

On September 24, 2018, The Wall Street Journal published internal emails showing that Google employees had held a group discussion about adjusting search features in response to President Trump's travel ban, so as to tell users how to donate to pro-immigration organizations and how to reach lawyers and government agencies. Google has since maintained that it does not manually intervene in particular search results.

There is no evidence that Google or Facebook controls the flow of information online out of malice. But on the enormous influence they wield over how information travels, America's two political camps have started from opposite points and arrived at the same worry.

Google, accused of taking away media ad revenue, is seen as a main cause of journalism's decline, narrowing the range of information

One reason fake news flourishes is that traditional media has weakened, and that the way internet platforms present information makes it hard to judge how authoritative a source is.

In 2002, Google launched its news service. News gets more than 4 billion clicks a month: 1 billion from Google News itself, the other 3 billion from web search.

Getting the news came to mean searching on Google or looking at a social network. Since 2013, traffic sent by Google and Facebook has been more than 50% of all traffic to media websites.

The direct result is that the American media business, which depends on page views for its ad revenue, went into steep decline. In July 2018 the left-leaning Guardian ran a piece arguing that Google is strangling journalism and democratic freedom, on the grounds that Google has won away the ad revenue that used to belong to the news business.

Since 2005, total newspaper revenue has fallen from $50 billion to under $20 billion. The number of people working in American newsrooms has dropped from more than 400,000 in 2001 to fewer than 185,000 today.

Free discussion in the press is a right the First Amendment has protected from the beginning, and antitrust enforcement in both telecommunications and newspapers has aimed at keeping a plurality of voices in the news. The FTC once issued a policy protecting newspapers, allowing them to merge on the revenue side but requiring that two separate papers and editorial teams be kept.

The tech giants' answer has been to fund the media themselves. Over the past few years Google has kept up its support for local news. In March 2018 it launched the Google News Initiative (GNI), a program to support news websites, under which Google planned to invest $300 million.

The internet's founding vision of information flowing freely has given way to compromise and censorship, and nobody has found a better answer

In August 2018, The Intercept reported that Google was building Dragonfly, a search engine with special censorship functions.

Protest came from every direction. Both parties condemned Google, and a group of 16 lawmakers from both sides quickly formed to find out what was going on. The pressure from inside Google mattered as much as the pressure from outside. Within a week, 740 engineers had signed a protest under their own names. On December 12, Pichai took questions about the project at a hearing, answering with the same sentence over and over: "Google has no plans to launch [the special-edition search engine] at this time."

It was not the first choice of this kind Google had made. In 2016, Thailand's military government said Google had agreed to its request and removed YouTube videos deemed insulting to the Thai royal family. Google publishes transparency reports listing removal requests from governments around the world, but the fact that it has not refused such requests still draws considerable criticism.

Earlier, in 2012, Google blocked access in Egypt and Libya after an anti-Muslim video set off violence there. Kevin Bankston, then head of the Free Expression Project at the Center for Democracy and Technology, described Google as walking a dangerous tightrope: on one side, blocking a video harms free speech; on the other, content on the platform that incites violence has real social consequences.

This is a question the tech giants, Google included, ask themselves. Jason Pontin, former editor of MIT Technology Review, was commissioned by Google to study one question: can it protect free speech and police harmful content at the same time?

When the internet was born, the people pushing it forward hoped for a new world where anyone, anywhere, could express their beliefs. As EFF founder John Perry Barlow put it in his 1996 "Declaration of the Independence of Cyberspace": "We are creating a world that all may enter without privilege or prejudice accorded by race, economic power, military force, or station of birth. We are creating a world where anyone, anywhere may express his or her beliefs, no matter how singular, without fear of being coerced into silence or conformity."

That declaration promised that the internet would let the dreams of Jefferson, Washington, Mill, Madison, Tocqueville and Brandeis be remade and come alive again through the internet.

The internal research Pontin did for Google reached a different conclusion. Google needs to do more moderation if it wants to please governments around the world and keep expanding globally. The data also shows that doing so pleases advertisers — the largest force behind the free commercial internet. Judging by what has happened since, that report's conclusion has become the internet giants' working principle.

In 20 years the tech giants have gone from a string of young startup stories to switches embedded in every corner of society, controlling the flow of information, shaping — deliberately or not — what billions of people buy, what they watch, how much privacy they keep, even how they work.

Now the harms of those switches are being laid out one by one, and there is no obvious conclusion about how to deal with them.

Some governments have started to act. Recent cases include Singapore legislating against fake news on online platforms and Britain issuing its "Online Harms" white paper, which sets out explicit penalties for internet companies that moderate poorly. Earlier still, many countries passed their own versions of an internet security law, most of which give the government sweeping powers to intervene.

But even from the accusations aimed at one company, Google, it is not hard to see that direct intervention by public power often creates a bigger problem.

Cover image source: Wiki