|Worth posting in full:|
Everyone’s Mad at Google and Sundar Pichai Has to Fix It
The CEO is increasingly boxed in by regulators, tech critics on both the right and the left, and even his own employees.
By Mark Bergen and Brad Stone
October 23, 2017
Photographer: Finlay MacKay for Bloomberg Businessweek
The sparsely furnished offices of Sundar Pichai stretch across the second floor of the Googleplex in Mountain View, Calif. On one side of the room, a sofa and chairs surround a coffee table bearing a few figurines, including a wooden dinosaur, the unofficial mascot of the Chrome browser. Dominating the other side is a massive treadmill desk, though Pichai rarely uses it. “I find it difficult to walk and type emails at the same time,” he says. “I’m not good at multitasking.”
That’s a problem, because being chief executive of Google lately has pretty much required world-champion grandmaster multitasking skills. Pichai has to run the world’s second-most-valuable company while managing political attacks and cultural blowups that seem to arrive every week. Since he was appointed Larry Page’s successor two years ago, he’s had to deal with a staff protest over the president’s immigration policy, a prolonged standoff with advertisers over unseemly videos on YouTube, a record regulatory fine, debates about gender inequality, and a growing sense around the globe that the tech giants—Google chief among them—are too big, too powerful, and perhaps too careless with the trust that their billions of users have invested in them.
Then, of course, there’s the matter of fake news. In the past few months, investigators have homed in on Google, Facebook, and Twitter for the role they played in the confusion about what was and wasn’t real during an election that may have been swayed by a foreign government. Google, like its Silicon Valley brethren, has turned over evidence to federal investigators that Russian interlopers bought political ads on YouTube, AdWords, and other of its services last year; representatives from each company will testify before Congress on Nov. 1. “There’s clearly stuff which shouldn’t be happening which happened, so we should fix it,” says Pichai, sitting in one of the chairs across from the sofa, which he calls the “psychiatrist’s couch.” He sounds a little like he should be lying on it himself. “Anytime we make a mistake, it’s very public for the world to see.”
A native of Chennai, India, Pichai joined Google in 2004. For a decade he stood out as a loyal and patient product manager, overseeing some of the company’s most popular services, such as Gmail, the Chrome browser, and Android. In 2015 his bosses, Page and Sergey Brin, formed Alphabet Inc., a new parent company, and focused their own energies on futuristic endeavors such as self-driving cars, internet transmission balloons, and life extension. They left Pichai in charge of the search and advertising business at Google, which generates more than 99 percent of Alphabet’s revenue and all of its profit.
Pichai has used that opportunity to recast Google’s mission in the most dramatic way since the search engine went live 20 years ago. Inside and outside the company, he’s elevated the role and glorified the promise of artificial intelligence—the ability of advanced computers to make independent decisions. Those decisions might be as small as when to flag a calendar appointment, or as consequential as how a multibillion-dollar hedge fund might trade. Either way, AI tends to evoke strong responses from people. To alleviate public anxieties about AI being just another way to kill jobs, Pichai recently embarked on a goodwill journey to the Rust Belt. “We understand there’s uncertainty and even concern about the pace of technological change,” he told an audience in Pittsburgh on Oct. 12, while announcing a digital training initiative and a $1 billion grant to nonprofits that retrain workers.
In Pichai’s office, conversation turns to the tech industry’s role in the proliferation of fake news. For example, on the day of the mass murder in Las Vegas, as reports came in that 58 people were killed and hundreds injured, prominent news headlines on Google and Facebook falsely pegged the killer as a Democratic opponent of Donald Trump. Discussing the issue, Pichai deploys the vocabulary of an apologetic CEO that’s become de rigueur in Silicon Valley since last November. He says the word “thoughtful” 13 times and “deeply” (feeling, listening, engaging…) six times, and proclaims the need to “do better” five times.
Page and Brin, Google’s founders, could often come off as Vulcans—cold-hearted calculators who used an engineering mindset to approach the world’s problems. Pichai, though, has the empathy and introspection of Captain Picard (and the beard of Commander Riker). In fact, his likability may be Google’s best weapon in its mounting political and business fights. “To have someone who exudes sincerity and is not hubristic is a comparative advantage at this time, when a little bit of humility is in order,” says Robert Thomson, CEO of News Corp., a persistent Google foe.
Dave Girouard, a former Google executive, puts it another way: “Good God,” he says. “They got the right guy at the right time.”
On a Monday last January, thousands of Google employees walked out of the office for a rally sanctioned by the company. They were protesting President Trump’s executive order banning travel from seven predominantly Muslim countries. They were also venting the anger that had spread since the election throughout Silicon Valley and among Google’s rank and file, which is populated by top computer scientists and engineers from around the world.
Dressed in a purple hoodie and sunglasses, Pichai spoke to the staff, summoning his own story as an immigrant and reassuring them that management, too, was wrestling with the implications of the ban and how to counter it. Even though he’s now a figure of authority, he managed to harness employees’ anger rather than become a target of it. He finished with this, as the crowd chanted his name: “We all need to learn to reach out and communicate to people from across the country. The fight will continue.”
Unlike Alphabet’s executive chairman, Eric Schmidt, a partisan fixture who backed Hillary Clinton, Pichai usually keeps his politics to himself. He frames his actions less as ideological than paternalistic, an effort to protect his 70,000-employee flock. “I could feel the company’s pain,” he recalls. “I felt it was my calling to do it.”
With the speech, Pichai earned some relative peace—for about six weeks. Then reporters for the Times of London found ads for big brands such as L’Oréal parked next to YouTube videos with jihadi and neo-Nazi messages. Whoops. The newspaper contacted the advertisers, who promptly pulled their ad dollars from YouTube. At first Google downplayed the controversy, but it responded after the boycott spread across the Atlantic and global brands such as AT&T and Johnson & Johnson scurried away from any connection with online hate.
The episode demonstrated the limits of Google’s technology; its vaunted AI was unable to spot hateful or profane material. Pichai’s defense is, essentially, that the problem is much more difficult than it looks. “Where do you draw the line?” he asks. “Sometimes, depending on what the world values at the moment, you have to redraw that line right then.”
That may be true from a computer science perspective, but it’s not terribly reassuring to the companies that pay Google’s bills. At a private meeting with anxious ad agencies in September, a Google manager tried to demonstrate the complexity of the issue by referencing a massively popular YouTube music video depicting a violent bank heist, from the artist The Weeknd. Thirty seconds in, a robber pulls out a gun and shoots a security guard in the head. More carnage ensues. The Google manager argued that the video was fictional and comparable to shows on traditional TV, according to two people there. Some advertisers vocally disagreed; their clients would rather not advertise baby products next to high-production-value depictions of homicide. Google’s “response primarily through the whole event was, ‘It’s really hard,’?” says one participant, Rob Bernstein, managing director of Reprise Media.
Pichai’s talk-it-out approach also faces challenges in Europe. In June the European Union levied a $2.7 billion fine against Google in the first of three antitrust cases, forcing the company to adjust its $79 billion-a-year ad business. The fine was expected, but Google was caught off guard by a fresh onslaught of accusations of self-dealing in search results. (For the past few years, Google has supplied its own answers to searches for products. People save time but click to other websites less often.) Still, Pichai seems to think he can mollify European Commissioner for Competition Margrethe Vestager and any trustbusters that come his way. “I generally feel like if people talked together more and engaged with each other, it leads to better outcomes,” he says. He adds that the company is “committed to finding a solution which the European Commission is happy with.”
Google is still appealing the EU decision, but the static across the Atlantic is another example of how Pichai is increasingly boxed in—by regulators, right-leaning critics, left-leaning critics, and sometimes even his own employees. In August he flew to Nigeria and Europe for Google events, then met his family for vacation in Portugal. Travel is often difficult for Google’s boss; his assistants usually bring along a supply of ginger cookies and candies—his preferred way to quell motion sickness, according to a former employee. This time, Pichai would need something stronger.
Shortly after landing in Portugal, he learned about The Manifesto. James Damore, an engineer in Google’s search division, had circulated a document earlier in the summer titled “Google’s Ideological Echo Chamber” that criticized the company’s treatment of conservative points of view and its efforts to hire more women in engineering. On Aug. 2, a Wednesday, Damore forwarded the note to an internal email list called “skeptics.” The next day, Google employees, aghast at his misogynistic line of argument, began tweeting about the memo. By Saturday it was public, and Pichai was on a flight back home to address a growing crisis.
The Damore memo set up an agonizing choice that positioned two of Google’s most important values—freedom of expression and employee harmony—against each other. Damore had a right to his opinion but had enraged both women and men at Google. Back in the Bay Area, Pichai met with John Hennessy. A former Stanford University president who sits on Alphabet’s board, Hennessy has become a mentor to many Google executives after the passing last year of longtime adviser Bill Campbell. Hennessy says the episode “blew up faster than we thought,” but that it was clear what needed to happen. “The way he wrote that memo—basically, a lot of women at Google felt like he was saying, ‘You don’t belong here,’?” he says.
On Aug. 7, Pichai fired Damore for violating Google’s code of conduct. “I was making the decision in the context of Google as a workplace,” he says. “But I realized it would have impact beyond it.”
It did. Conservatives lionized Damore as a culture warrior against Big Tech and political correctness. Pichai had to cancel a company all-hands meeting to discuss the drama after employees reported receiving online threats. Tucker Carlson ran several segments on his Fox News show asking if Google was too big. Posters appeared along Los Angeles sidewalks showing Steve Jobs and Apple’s famed motto, “Think Different”; below was Pichai’s bearded face and the phrase “Not So Much.” David Brooks piled on: “Sundar Pichai Should Resign As Google’s C.E.O.,” read the headline of his column in the New York Times.
Pichai probably didn’t figure that defusing political and cultural grenades would be such a big part of the job. But while some critics charged that Google waffled on the issue and let it boil over into a PR nightmare, others saw a surer hand in Damore’s firing. Scott Galloway, a New York University professor and author of The Four, a critical book about big technology companies, compliments Pichai’s response. “It was a crisp, clear decision; he made it fast,” he says. “Google would be the scariest company in the world if you didn’t believe they had adult supervision.”
Pichai’s solution to the gnawing problems of fake news and illicit content that slip through Google artificial intelligence is, no surprise, more artificial intelligence. He thinks humans will succeed in training AI and that it’ll all be worth the effort. Like other technologists, he believes AI can do far more for both Google and the world, dramatically improving transportation, health care, agriculture, and any other field that uses computers. And he’s reshaped the company on the premise that the age of AI will usher in opportunities that dwarf even the size of the internet economy.
Pichai traces his enthusiasm for AI to the end of 2014, when a team at Google was itching to release an enhanced version of Google Photos. The app used eerily accurate image-recognition tools, automatically tagged pictures, and organized them based on the people who appear in each shot. Pichai, then head of all company products, decided to introduce the app the following May at I/O, the company’s annual conference for developers. There he performed what’s now a staple of his CEO stagecraft: some deeply technical musings about AI and its implications. In that speech, he spent a full minute in geek mode, explaining how Google’s Al-flavored algorithms can spot a tree frog in a photo, thanks to “deep neural nets” and mathematical models forming a “hierarchical learning system” that’s “30 layers deep.”
His minders didn’t love the wonky stuff, but Pichai says he insisted. The launch was what he calls the “crystallizing moment” for the new narrative—that Google would now be an AI-first company. “I felt if you showed how the technology works to make something better, tangibly better for people,” he says, “it’s a big thing.” It was so big that Pichai devoted all of the company’s annual memo to investors to the topic of how AI can do everything from accomplish daily tasks to, one day, make cancer diagnoses and solve climate change.
And yet, so far, some of the more prominent AI at work in Google’s consumer products doesn’t exactly impress. In May, for example, the company introduced a feature called Smart Reply, which assesses the emails of harried Gmail users and furnishes canned answers (“I’m in!”). That’s a bit sad because, at best, the feature risks saving users a few seconds while cheapening the quality of their online correspondence. Pichai, naturally, doesn’t see it that way. He muses about weaving AI into the Android operating system so that users might get an alert for an urgent calendar appointment—from their doctor, for example—even if they’ve blocked notifications on their phones. “You know,” he says, “Smart Reply in Gmail is actually a very difficult thing to do.”
Inside Google, Pichai’s biggest decision has been to set up two divisions—cloud computing and hardware—with a goal to channel AI into significant technical advances. In the cloud business, his bet is that AI-equipped tools can set Google apart from its main rivals, Amazon.com Inc. and Microsoft Corp. Google recently began offering cloud access to specialized chips designed to enable machine learning. The hope is that even ancient industries can harness the technology’s magic. For example, Connecterra, an Amsterdam-based startup, uses Google AI to create digital sensors that farmers can use to monitor their livestock’s movement and eating habits. It may take a while for Pichai’s strategy to pay off, though. Connecterra actually pays Microsoft to rent cloud storage, the main way Google hopes to cash in on its AI tools. And Amazon still dominates the cloud services market, with sales 12 times greater than Google’s, according to an estimate by Raymond James & Associates. “AI cloud is just very, very nascent,” argues Fei-Fei Li, director of the artificial intelligence lab at Stanford, whom Pichai hired as chief scientist for the division.
In early October, at an event in San Francisco, Pichai unveiled the Pixel 2, Google’s latest smartphone, and a line of other devices with built-in AI smarts. The highlight was the Pixel Buds, a pair of wireless headphones that can help users translate and speak in foreign languages on the fly. It’s the realization of a long-held sci-fi dream, embodied by the universal translator in Star Trek or the Babel Fish in the books of Douglas Adams. If it works, it’s a product that finally fulfills AI’s promise. “Computing needs to evolve to where this happens in a more natural, seamless, and conversational way,” Pichai said at the event.
The company also introduced versions of its Google Home voice-activated speaker. But Amazon appears to have the edge, and not only because it was first to market, or because Alexa, its own AI assistant, performs charmingly well. Amazon also prices its speakers cheaply and recoups its investment by turning device buyers into frequent shoppers.
Google has historically stumbled in the difficult business of hardware, despite several multibillion-dollar attempts, including the $12 billion acquisition of Motorola Mobility (since sold off for $2.9 billion) and the $3.2 billion purchase of smart-appliance maker Nest Labs (a division that’s widely seen as struggling). In September, Google tried again, spending $1.1 billion for around 2,000 engineers from HTC Corp., which assembled the Pixel phone. That deal had been “on the back burner” for some 18 months, says Rick Osterloh, Google’s hardware chief, until he and Pichai decided to proceed.
The Pixel 2 and Pixel Buds drew praise, but Google, being Google, somehow managed to mar the moment. Six days after the hardware unveiling, a tech blogger noticed his review unit of Google’s miniature speaker was surreptitiously recording everything in his house. Google blamed the bug on a new feature that allows users to activate the device with a tap, then killed the tap function.
Another new device, Clips, is a small camera that automatically snaps and sorts photos on its own. “This doesn’t even *seem* innocent,” tweeted Elon Musk, one of Google’s most steadfast adversaries. Musk, who was suggesting the device was snooping on users, has stirred general alarm about AI; he thinks computers that make their own decisions and are smarter than people could enslave humanity.
During the interview at the Googleplex, Pichai’s face lights up at the mention of Clips, then darkens when he hears about Musk’s barb—he hadn’t seen it. “He’s a deep thinker about problems, and I think it’s right to be concerned about AI,” he says, with typical diplomacy. Advances in the tech can’t be stopped, so Google’s role, he argues, is to be a “thoughtful and ethical” steward of the technology. “None of us have answers yet, which is what Elon is pointing out.”
If Elon Musk wanted to kvetch directly to Google, he’d probably go directly to his old philosophical sparring partner, Larry Page. Which brings up the main thing that sets Pichai apart from most of his CEO peers. It’s not his humility, diplomacy, or enthusiasm for AI. It’s that he has a boss.
Page always managed to delegate the political and managerial messes that Pichai now has to confront. Page is also largely invisible to the media; his last interview was two years ago. (He declined to speak for this story.) Yet he and Brin still control 51 percent of voting shares in Alphabet. One executive who recently left Google describes Pichai’s role more as a chief operating officer; another equates it with working for a “family company.”
Hennessy, the Alphabet board member, contests any suggestion that Pichai isn’t in charge. “Sundar would take any big decision to Larry just to get his views,” he says. “But Sundar is really running Google, and Larry has doubled down on the Other Bets”—company jargon for Alphabet units that aren’t called Google. Waymo, which works on self-driving cars, is an Other Bet, as is Verily, which does biotech. Pichai says he meets with Page about once a week and can call anytime he needs. “The biggest thing Larry does for me is to always help me take a long-term perspective,” he says.
What’s clear is that Pichai is taking on long-standing problems that Page, in all his visionary brilliance, never much cared about. For years, to cite one of many examples, Google had tenuous, often hostile relations with news publishers, who watched with seething rage as their advertising dollars were vacuumed up by digital distributors who paid nothing to produce actual content. Thomson, the News Corp. CEO, endlessly voiced those gripes to Google only to feel like a rejected character from a Charles Dickens novel. “If you waited for the phone to ring” back, he says, “you’d be Miss Havisham in her wedding dress all those years.”
But Pichai returns phone calls. He’s working with News Corp. and other media companies on subscription tools and has scrapped a rule demanding free articles in search results, which punished news organizations that tried to charge for professionally produced work.
Pichai is also confronting the problem of fake news and less credible news sources in a way Google usually doesn’t do: by adding more humans. Earlier in the year he appointed a “fake news czar,” as some call a new product manager for Google News. Pichai says he’s also considering promoting vetted news sources in Google’s News search results—elevating, say, articles in the Las Vegas Review-Journal over posts on the anything-goes bulletin boards 4chan or Reddit Inc. “We want to be balanced in making sure a range of perspectives are available, but I think a flight to quality is important,” he says. “We can’t make mistakes.”
It’s a bland statement—but coming from a Google executive, it’s stunning. Silicon Valley executives have generally been dismissive of traditional media outlets and have operated under the assumption that it’s better to move fast, break things, and, unless absolutely necessary, not ask for forgiveness. But times have changed. As Pichai has learned, Big Tech companies can no longer skate by on faith in their fundamental benevolence. —With Ellen Huet