In a San Francisco federal courtroom, a former Meta engineering director has offered testimony that places one of the world's most powerful technology companies at a moral crossroads it long chose to avoid. Arturo Béjar's account suggests that the architecture of harm visited upon millions of children was not accidental but structural — built into the logic of growth itself. As 29 states seek accountability in what may become the defining legal reckoning of the social media era, the trial asks a question older than any algorithm: what do we owe the most vulnerable among us, and who decides?
Meta Engineer Testifies Zuckerberg Prioritized Growth Over Child Safety
Mountains move in months when Mark makes something a priority.
So Béjar worked at Meta for years. What made his testimony different from what we might expect from any disgruntled former employee?
He wasn't just complaining about management. He testified about specific design choices—the "don't ask, don't tell" policy, the deliberately ineffective "take a break" feature. These were decisions, not accidents.
But we should note: Béjar was not working directly with Zuckerberg during his second stint at the company. He acknowledged that. So when he says Zuckerberg made these choices, how much of that is direct knowledge versus inference from the culture he observed?
Fair point. But he also testified before the Senate in 2023 about the same patterns, and he's now appeared as a key witness in four trials. The consistency matters.
What about the "don't ask, don't tell" policy? That's a pretty specific allegation.
Béjar said Meta had the technology to flag underage users but didn't use it because identifying them would reduce profits. That's the core claim—not that they couldn't protect children, but that they chose not to.
And that's based on his testimony and understanding of the company's incentives. We should be clear: we don't have internal Meta documents saying "let's not flag underage users to protect profits." We have Béjar's interpretation of why the policy existed.
What about the "take a break" feature? That seems like something Meta could point to as evidence they were trying to help.
Béjar said it was designed to fail—not the default, easy to ignore. And Meta's own researcher testified the company could make Facebook less habit-forming but didn't.
That's from Elena Davis's recorded testimony. She was reading from a report. We should know: was that report commissioned by Meta, or by the plaintiffs? The source of that research shapes how we read it.
What's the human cost here, beyond the legal numbers?
Béjar's daughter is one example. Harassment, misogynistic abuse. But the testimony also mentioned a content moderation team reviewing 15,000 posts monthly related to child sexual abuse. That's the scale of what's happening on these platforms.
That's a real number from the source material. But we should ask: does that 15,000 figure represent the total volume of such content, or just what one team reviewed? The source doesn't clarify.
Where does this go from here?
Six weeks of trial. Judge Gonzalez Rogers will decide liability and penalties. The outcome could reshape how social media companies approach child safety.
Or it could result in a settlement that doesn't change much. We won't know until the verdict comes in.
Il Polso
- A former insider testified that Mark Zuckerberg's command-and-control culture meant child safety improvements only moved when he personally willed them — and too often, he did not.
- Meta allegedly deployed a 'don't ask, don't tell' approach to underage users, deliberately avoiding identification of children under 13 because doing so would have cost the company long-term revenue.
- Béjar's own teenage daughter encountered misogynistic harassment on Instagram while searching for community, giving his testimony a personal weight that no corporate rebuttal could easily absorb.
- Meta's defense pointed to its hundreds of safety employees and gaps in Béjar's direct access to Zuckerberg, but internal research from a company researcher suggested Meta knowingly chose engagement over reform.
- With potential penalties ranging from $200 billion to $1.4 trillion and prior jury verdicts already finding Meta liable, the six-week trial is rapidly becoming the most consequential legal test of Big Tech's duty of care to children.
In a San Francisco federal courtroom, a former Meta engineering director has offered testimony that places one of the world's most powerful technology companies at a moral crossroads it long chose to avoid. Arturo Béjar's account suggests that the architecture of harm visited upon millions of children was not accidental but structural — built into the logic of growth itself. As 29 states seek accountability in what may become the defining legal reckoning of the social media era, the trial asks a question older than any algorithm: what do we owe the most vulnerable among us, and who decides?
On Wednesday, Arturo Béjar — a former engineering director who spent years inside Meta's product teams — took the stand in federal court and described a company where child safety consistently yielded to the demands of growth. Testifying before Judge Yvonne Gonzalez Rogers, Béjar said that meaningful change at Meta required Zuckerberg's personal blessing: when the CEO cared, he said, "mountains move in months." When he did not, nothing moved at all.
The trial, led by California, Colorado, Kentucky, and New Jersey on behalf of a 29-state coalition, opened Tuesday and is expected to last six weeks. The states allege that Meta deliberately engineered Facebook and Instagram to addict young users, misled the public about safety measures, and violated federal law by harvesting data from children under thirteen. Meta has denied the allegations in full.
Béjar's testimony struck at the heart of statements Zuckerberg made in 2021, when he publicly dismissed whistleblower claims that the company knew its platforms were harming children. "You just cannot trust Mark Zuckerberg with kids," Béjar said flatly. He spoke with personal authority: his own teenage daughter had joined Instagram hoping to connect with other women who loved cars, only to be met with a torrent of misogynistic comments about her body.
He also described Meta's "take a break" safety feature as designed to be ignored — not set as a default, easy to dismiss — and testified that the company possessed technology capable of identifying millions of likely underage users but chose not to deploy it. Flagging those children, he said, would have reduced long-term profits, so Meta adopted a deliberate policy of looking away.
Meta's attorneys challenged Béjar's access and scope during cross-examination, noting he had not worked directly with Zuckerberg in his later tenure and that the company employs hundreds of dedicated safety professionals. But recorded testimony from Meta researcher Elena Davis described internal findings showing the company could have redesigned its platforms to be less compulsive — and chose not to.
The trial arrives amid a wave of legal pressure. A New Mexico jury earlier this year ordered Meta to pay $375 million for misleading users about child safety — the first time a court held the company liable for endangering children. A Los Angeles jury found Meta and YouTube liable for addictive design. With Meta's own estimates placing potential exposure as high as $1.4 trillion, the proceedings represent the most consequential reckoning yet between social media's architects and the children they shaped.
A former Meta engineering director took the stand Wednesday and painted a portrait of a company where child safety was treated as a secondary concern, subordinate to the relentless pursuit of growth and user engagement. Arturo Béjar, who spent years inside Meta's product teams, testified that the company's decision-making flowed downward from CEO Mark Zuckerberg with little room for dissent. When Zuckerberg decided something mattered, Béjar said, "mountains move in months." When he did not, the machinery of change simply stalled.
The trial opened Tuesday in federal court before Judge Yvonne Gonzalez Rogers and is expected to run for six weeks. Four states—California, Colorado, Kentucky, and New Jersey—are leading the charge on behalf of a coalition of 29 states that filed suit in 2023. The allegations are sweeping: Meta deliberately designed Facebook and Instagram to addict young users, misled the public about safety protections, and violated federal law by collecting and using personal data from children under thirteen. Meta has denied all of it.
Béjar's testimony directly contradicted statements Zuckerberg made in October 2021, after whistleblower revelations suggested the company knew its platforms were harming children but had chosen not to act. "It's so false, every part of it," Béjar said from the witness stand. "You just cannot trust Mark Zuckerberg with kids." He spoke from experience. Years earlier, when his own sixteen-year-old daughter set up an Instagram account hoping to find a community of women interested in cars, she instead encountered a barrage of misogynistic comments—remarks about her body that left her visibly distressed. She had gained followers, Béjar testified, but at the cost of real harm.
The company's signature safety feature, a tool called "take a break," was designed to fail, Béjar argued. It was not the default setting, and the reminders users received were easy to dismiss. Meta's own research, he said, had established that most parents lacked the time or knowledge to understand how their children actually used these platforms. Yet the company did nothing to change that equation. More damning still, Béjar testified that Meta possessed technology capable of flagging potentially millions of users who appeared to be under thirteen—but the company adopted what he called a "don't ask, don't tell" policy. Identifying those underage users would have reduced long-term profits, so the company simply looked away.
Meta's legal team pushed back during cross-examination. Béjar acknowledged he had not worked directly with Zuckerberg during his second stint at the company, from 2019 to 2021, and that he lacked access to certain internal resources. The company also noted it employs hundreds of safety-focused staff members, people Béjar himself considered qualified. But the broader pattern of testimony painted a different picture. A Meta researcher named Elena Davis provided recorded testimony describing internal reports showing the company could redesign Facebook to be less habit-forming, yet chose not to.
This trial represents the most significant legal reckoning yet with social media's effects on children. Meta faces thousands of similar lawsuits. Earlier this year, a New Mexico jury ordered the company to pay $375 million for misleading users about child safety—the first time a court had held Meta liable for endangering children. A Los Angeles jury also found Meta and YouTube liable for designing addictive platforms. The company's own estimates suggest potential penalties could reach $1.4 trillion, though state attorneys general have suggested a figure closer to $200 billion. The stakes are enormous, and the testimony from engineers and researchers who worked inside Meta's walls suggests the company understood the risks all along.
Citazioni salienti
If Mark makes something a priority, mountains move in months.— Arturo Béjar, former Meta engineering director
It's so false, every part of it. You just cannot trust Mark Zuckerberg with kids.— Arturo Béjar, disputing Zuckerberg's October 2021 statements denying knowledge of child safety risks