Why Arturo Béjar left Meta – and blew the whistle: ‘I don’t recall Mark ever being empathetic’
The Social Reckoning hits cinemas this Friday, but the real world has already been working on a sequel – or rather, a threequel. Directed by Aaron Sorkin, the movie is a follow-up to The Social Network, detailing the further misadventures of Mark Zuckerberg and his social media empire, and in particular Frances Haugen, the employee who in 2021 leaked internal documents that appeared to show Facebook was aware of the harms its products had inflicted.
It’s no spoiler to reveal the film climaxes with Haugen’s testimony to the US Senate, on 5 October 2021. That very same day, Arturo Béjar, a consultant on the wellbeing team of Instagram (which is also owned by Facebook), hit send on a long-gestating email to Zuckerberg and other senior Facebook executives on a similar theme. “I wanted to bring to your attention what I believe is a critical gap in how we as a company approach harm, and how the people we serve experience it,” Béjar began, before laying out some shocking statistics from his team’s survey. For example: 24.4% of 13-15-year-olds on Instagram said they had received unwanted advances, and 21.8% said they had been the target of bullying on the platform in the past seven days. Furthermore, 51% of Instagram users said they had had a bad or harmful experience in the past seven days. Only 1% of them reported it, and of those, 2% had the content taken down – so just 0.02%. This was a far greater degree of harm than the company’s existing measurements suggested. Zuckerberg never replied.
By that time, Béjar had worked at Facebook (which rebranded as Meta later that month) for a total of eight years. He had learned how to communicate with Zuckerberg. “You couldn’t talk to Mark about, ‘We have to do this thing to prevent suicide because it’s the right thing to do,’” he says. “You had to put everything in terms of metrics. That was his language. That’s how he related to things, which is also what’s behind all the grief in the world right now.”
Béjar has seen The Social Reckoning, and he recognises Jeremy Strong’s portrayal of Zuckerberg as a detached, awkward, humourless control freak all too well. “Oh, it’s accurate,” he says. “I don’t recall him ever being empathetic to the harm that somebody was experiencing and wanting to make that better … The only thing that really consistently came across in all my interactions with him was winning. He’s one of the most fiercely competitive people I know.”
One time, Béjar recalls, he and Zuckerberg were walking together after a fraught meeting. “One of the things that I used to say when bad things were happening is, ‘Are we having fun yet?’ Just as a little kind of defusing humour thing,” says Béjar. “And he looked at me, very intense, and said, ‘I’m not here for fun – I’m here for impact.’ That was my experience of him, over and over and over again.”
That 2021 email was the beginning of the end for Béjar. He was still hoping Zuckerberg would respond to his appeal to engineer a “culture shift” in the way Instagram dealt with protecting its young users. The next day, however, Zuckerberg issued a statement to Facebook employees about Haugen’s testimony and the leaked “Facebook Files” that were being reported by the Wall Street Journal and others. “At the heart of these accusations is this idea that we prioritise profit over safety and wellbeing. That’s just not true,” Zuckerberg wrote. But by then Béjar knew it was true.
By 2023, Béjar was following in Haugen’s footsteps and testifying to the US senate himself. He has testified under oath against Meta at least seven times since, by his reckoning, in various trials and hearings. The most recent, and consequential, was the landmark trial in California in August, brought by the attorneys general of 29 US states, which alleged once again that Meta knowingly designed products that were addictive and harmful to minors. In his testimony, Béjar stated: “You just cannot trust Mark Zuckerberg with kids.”
In many respects Béjar is the perfect witness: sincere, calmly spoken, emotionally literate, able to express complex ideas in accessible ways, and possessing reams of inside knowledge. He’s seemingly the opposite of Zuckerberg. At times during our conversation, as we sit outside a cafe in north London, he has to pause mid-sentence, his eyes welling with tears. He apologises afterwards for getting emotional, but that seems to be exactly what these issues demand.
There’s another reason Béjar has been such a powerful spokesperson: he has a direct connection to the issues, via his teenage daughter. When she was in her early teens, in about 2018, she told him she regularly received sexist comments and unsolicited nude photos from men via her Instagram account. Her female friends experienced the same thing. There was no point reporting it, she told him, because it never resulted in any meaningful action. Commenters might get blocked – after which they would simply open a new account under a different name – or Instagram would judge that the behaviour didn’t violate their policies.
Until that point, like many early Facebook employees, Béjar believed he was making the world a better place. When he started in 2009, the company was booming and the idealism was infectious. “The promise was, we’ll create something that keeps you close to your friends. I think a lot of people back then believed in that. And that’s not what they built.”
Béjar, 55, was an early tech enthusiast. Born in Mexico City, the son of a doctor father and an artist mother, he taught himself to code and blagged his way into a job with IBM when he was 15. He came to Silicon Valley, after a spell in London, in 1993, and worked at Yahoo for 11 years before he joined Facebook.
As senior engineering and product leader at Facebook, Béjar was responsible for security but also for customer care: “developing tools that help people learn to be with each other online” – especially young people. For example, he developed a more nuanced way for young people to report content they didn’t like. If the button on Facebook said “report”, teens tended not to use it, but if it was changed to “I don’t want to see this”, three times as many teens would click on it. And if you could grade the intensity of that feeling, from “mildly annoying” to “really bad”, that made it easier to deal with problems.
“Most people who behave aggressively online, if you give them private, respectful feedback, they’re like, ‘Oh sorry, I didn’t mean to,’” Béjar says. “And that’s how we work as a society, right?” He gestures to the pedestrian street around us; people are peacefully going about their business. “What makes this street nice is not that there’s police everywhere enforcing our behaviour, it’s that we have all these ways where we signal to each other what it’s like to be safe, even though you might have fundamentally different values.” He built more than a dozen Facebook products on that premise, he says, with extraordinary results. “Basically, half the people change behaviour the first time you go, like, ‘Yeah, we don’t do that here.’”
He left the company, amicably, in 2015 because he was going through a divorce and wanted to spend more time with his young son and daughter. By the time he found out about his daughter’s Instagram experience, other young victims of social media were becoming known. In 2017, 14-year-old Molly Russell died from an act of self-harm in her London bedroom. The coroner’s report said her depression was exacerbated by her social media consumption, especially Instagram. Wider issues with Facebook were also beginning to emerge: its role in election manipulation, stoking ethnic violence, promoting inflammatory posts.
After raising his concerns with the company, Béjar was invited to return in 2019, on a two-year contract as a consultant on Instagram’s “wellbeing” team. “I still believed that the reason the stuff that was happening to my daughter and all her friends, and every other kid I’ve talked to since, was because as companies get big, sometimes the general doesn’t know what the soldiers are doing. That was my thesis for almost all the time I was there: Mark doesn’t know … and if he knew, he would fix it.”
Evidence to the contrary started piling up, though. “What I started finding when I came back is that they did know. They knew about addiction, they knew how bad it was. They knew about Molly Russell.” It later emerged that Facebook had internal research confirming Instagram’s negative effects on teenage mental health, but had kept it secret. The wellbeing team could not get anything done, Béjar says. “They could dream up a good product that would make things better, and they couldn’t get it out the door. You would put it through the review process and all the teams that would review it would just kill parts of it until there was nothing left.”
The ethos had changed, he says. For all the talk about connecting the world and bringing people together, the priority was simply growth, it seemed: attracting more users and keeping them on the platform for as long as possible. Video content was now set to autoplay; notifications were frequent, as were suggestions of other people to follow; “infinite scroll” algorithmically served up content tailored to users’ continuously tracked habits. All the customer care tools Béjar had previously developed – such as the “I don’t want to see this” button – had gone.
Béjar’s team developed a tool aimed at people with a history of posting comments that got reported and deleted – “the worst of the worst” – so that when they posted something inflammatory, they would get a community note telling them: “You might want to phrase this differently.” “We were in the process of getting that approved. And one of the people in the review meeting said, ‘What if 50 Cent gets this thing and makes fun of it? We can’t have that.’ That was the driver.”
The anecdote chimes with Facebook’s secretive “cross-check” system, revealed in the Facebook Files, where posts by celebrities, politicians and other high-profile “business partners” were moderated more leniently, even if they violated the company’s stated policies.
Had he known what he knows now, Béjar would never have allowed his daughter on Instagram in the first place, he says. “Instagram is built on three pillars: addictive design, connecting you to strangers, and rewarding you for doing things that are exhibitionistic, performative or aggressive,” he says. None of these are safe or desirable for young people, most people would agree. He compares social media to the tobacco industry: “This is the equivalent of handing your kids a cigarette and being like, ‘Hey, you’ll be cooler if you smoke.’” On unwanted sexual advances, he says, “What Instagram is effectively doing is, while your kid is in the bedroom, it’s lining up strangers outside the window and inviting them in.”
Béjar believes that the tone is squarely set by Zuckerberg, as the longtime CEO, chair and controlling shareholder of Meta: “If Mark woke up tomorrow and was like, ‘I want to create something that’s genuinely good for kids,’ it would take the company six months to fundamentally shift all these things and drive all these harms down. They have the technology, they have the people, they have the expertise, they have the infrastructure. It’s his refusal to do that which creates a culture.”
Ultimately, Béjar faced the same dilemma Haugen did. Speaking out meant ending his career, as well as being ostracised from his colleagues and his community. But, he says, “You have to ask yourself: who am I protecting? Am I protecting the company or my way of life or my being? Or am I protecting other people? It took me years, to really walk away from the identity that I had developed when I was working there.” He was emboldened by Haugen’s courageous example and by the treatment she received afterwards: “Internally, they would call her ‘the leaker’. They wouldn’t say her name.”
When approached for comment, a Meta spokesperson said: “We all share a common goal of providing safe, meaningful online experiences for young people. Our record shows that we’ve spent the past decade researching and developing numerous tools and resources to support teens and empower parents, including Teen Accounts, which place teens in protective settings by default. We’ll continue to make improvements and partner with parents and experts to keep teens safe.” Meta has repeatedly rejected implications that engagement stands in contrast to its commitment to teen safety.
Despite its best efforts to contrive a triumphant finale, The Social Reckoning ends on a bit of a sour note. There was no real reckoning for Meta. Donald Trump was re-elected; social media companies forgot about their commitment to fact-checking, Meta included. And Zuckerberg’s company continues to grow and profit from the digital economy – including the AI economy. Last month, still reeling from the reaction to Meta’s camera-enabled “pervert glasses”, Zuckerberg unveiled Muse, Meta’s cute and cuddly AI agent, which can perform personalised tasks such as shopping or sending emails. Even Haugen herself recently admitted: “We are worse off today than when I leaked the Facebook documents.”
Had they waited until this month, the film-makers might have been able to craft a happier ending. The landmark trial brought by multiple US states in August ended two days after Béjar’s testimony. Meta abandoned their challenge and settled the case (denying any wrongdoing). The company will have to pay out up to $18bn (£13.6bn); it’s a relatively small sum, considering that Meta’s revenue in 2025 was more than $200bn (£151.5bn). More significantly, for the first time, the company agreed to major changes to protect teens: two-hour daily time limits for under-18s across Instagram and Facebook; blocking notifications during school hours; blocking the apps at night; turning off autoplay; giving parents greater control; and other safety features.
“I think it’s an extraordinary accomplishment,” says Béjar. “Nobody else in the world has got them to agree that they will reduce the hours kids spend on the platform. Or that they’re going to be more transparent about how they assess age, or that it will reduce home notifications.”
These measures won’t fix everything, but they’re progress. Meta lost another trial, brought by the US state of New Mexico in March, again over misleading the public about safety, and had to pay $942m (£713.7m) in fines and damages. There are dozens more cases in the pipeline. Governmental regulation is tightening around the world. Perhaps the tide is turning.
If it is, it will be some relief to Béjar. He has spent much of the past three years campaigning, testifying, talking to governments, academics, regulators. He has other interests beyond the tech world. Even as he was quitting Meta, he was writing the libretto for a musical collaboration with the composer Philip Glass. He is writing an opera based on Charles Darwin’s theory of emotion in animals, with the British composer Jocelyn Pook. His daughter, by the way, now earns her living restoring classic sports cars.
There’s still work to be done. What’s really needed, Béjar argues, is independent assessment and comparison of the safety of social media sites – the same way we monitor, say, vehicle emissions. “Every parent and every kid deserves to know: what is the likelihood that this bad thing is going to happen to you on any of these services?” But at last it feels like things are headed in that direction, he says. “It’s still just the beginning, but I do feel hope.”
What's Your Reaction?
Like
0
Dislike
0
Love
0
Funny
0
Wow
0
Sad
0
Angry
0
Comments (0)