Meta Whistleblower Says Zuckerberg Put Growth Before Child Safety — Now He’s Testifying Under Oath
What has Arturo Béjar accused Meta of doing?
‘Don’t Ask, Don’t Tell’: Explosive Meta Testimony Puts Zuckerberg Under Pressure
One of the most serious accusations ever made about Meta’s treatment of children is no longer confined to leaked documents, congressional hearings or former employees speaking from outside the company. Former Meta engineering director Arturo Béjar has now testified in federal court that Mark Zuckerberg’s leadership created an organisation where growth and engagement repeatedly took precedence over protecting young users. Meta strongly disputes the allegations.
The stakes go far beyond whether one former employee is believed. Meta is facing a landmark case involving 29 US states over allegations surrounding children, addictive product design, underage users and representations the company made about the safety of Facebook and Instagram. The states are seeking potentially enormous financial penalties and changes that could reach directly into the architecture of the platforms themselves.
What Arturo Béjar Says Happened Inside Meta
Béjar is not a distant critic of Silicon Valley. He worked at Facebook between 2009 and 2015, including on safety systems, and later returned as a contractor working with Instagram between 2019 and 2021. His experience put him unusually close to the systems now being examined in court.
His central accusation is that protecting users from harmful experiences was structurally weaker than Meta's obsession with engagement. In testimony this week, Béjar described the company's approach to children under 13 as effectively “don't ask, don't tell”, alleging Meta could identify more underage users but had little incentive to pursue a system that might remove them from its platforms.
That matters because Meta officially requires Instagram users to be at least 13. The states' case goes considerably further, alleging Meta knew that large numbers of younger children were nevertheless using its services while their data was being collected. The original multistate complaint alleged that Meta's own information showed millions of under-13s on its platforms and accused the company of violating children's privacy protections.
Béjar has also described Zuckerberg as extraordinarily influential over what the company ultimately prioritised. His argument is not simply that individual safety employees made mistakes. It is that incentives flowing from the top placed immense importance on growth, time spent and engagement, meaning safety improvements faced a fundamentally different test when they threatened those objectives.
The Warning Sent Directly To Zuckerberg
Perhaps the most damaging element of Béjar's story is that he says senior management cannot plausibly claim it had no warning.
In written testimony previously submitted to the US Senate, Béjar described an internal analysis that found a profound difference between the amount of harm users said they experienced and the amount captured by Meta's conventional reporting systems. According to Béjar, 51% of surveyed Instagram users reported some form of negative experience in a week, yet only around 1% reported the offending content and only around 2% of those reporting it succeeded in getting it removed.
The figures involving younger teenagers were more troubling. Béjar told lawmakers that internal research initially suggested 21.8% of 13-to-15-year-olds surveyed had been targets of bullying during the previous seven days, while 39.4% reported negative social comparison. An initial figure of 24.4% for unwanted advances was subsequently revised by the research team to around 13%, according to his testimony.
On 5 October 2021, Béjar says he sent Zuckerberg and other senior executives a detailed email explaining what he had found.
Béjar testified that Instagram chief Adam Mosseri requested a follow-up meeting and that then-chief operating officer Sheryl Sandberg expressed sympathy after Béjar discussed what had happened to his daughter. Zuckerberg, he said, did not reply.
That episode has become important because the argument facing Meta is increasingly about knowledge rather than simply outcomes. Social networks containing harmful material is one question. Whether executives were shown evidence of recurring problems and then failed to make sufficiently substantial changes is a much more dangerous one.
His Daughter Changed The Story
The issue also became personal for Béjar.
He has testified that his 14-year-old daughter began receiving sexually inappropriate approaches from strangers after joining Instagram. In court this week, he described discovering requests for sexual material and explicit images sent by people she did not know.
For somebody who had spent years working on Facebook safety systems, the experience appears to have changed his understanding of the problem. His concern was not merely that bad users existed. It was that the machinery designed to connect people, recommend accounts and maximise interaction could also create pathways through which vulnerable young users encountered adults they did not know.
Béjar has previously argued that Meta's measurement of safety concentrated too heavily on whether content technically violated policies rather than asking users whether an interaction had actually harmed them. Something could therefore be deeply distressing without necessarily triggering the company's narrowest violation metrics.
That distinction sits at the centre of the wider controversy. A platform can remove enormous quantities of prohibited material and simultaneously leave millions of users experiencing interactions they regard as harmful. Both statements can theoretically be true.
Why Growth And Engagement Matter So Much
Meta's business model makes the argument especially consequential.
Facebook and Instagram are overwhelmingly advertising businesses. More activity creates more opportunities for advertisements to be shown, while products capable of retaining users are generally more commercially valuable. Engagement itself is therefore not evidence of wrongdoing; virtually every consumer technology company wants people to use its products.
The question in this case is where the commercial objective ended and the duty to protect children began.
The states' original complaint alleged that Meta internally maintained explicit targets relating to time spent on its services even while publicly rejecting the idea that simply maximising time was its objective. The publicly released complaint also alleged that the company closely monitored engagement among teenage users.
Béjar's testimony now gives that documentary argument a human witness. He says product teams understood what leadership valued and that child-safety interventions could struggle when they conflicted with engagement or adoption.
Those remain allegations. Meta denies designing its products to harm or addict children, and its lawyers argue that the states are trying to convert an enormously complicated debate about adolescent wellbeing into a simplistic story about a technology company deliberately causing harm.
Meta Has A Very Different Version Of Events
Meta's defence cannot be ignored because the company has introduced substantial child-safety measures over the years.
It has pointed to protections including restricted messaging, parental supervision, private-by-default accounts for younger users and Instagram Teen Accounts. Meta has said teenagers under 18 are automatically placed into Teen Accounts with additional protections, while those under 16 require parental permission to weaken certain settings.
The company has also directly challenged Béjar's interpretation of its internal surveys. In formal responses to US senators, Meta said the surveys he cited were intended to measure users' subjective experiences rather than the prevalence of content violating Meta policies. It said the information nevertheless informed its safety work and argued that its automated systems proactively detected large proportions of bullying and harassment content it ultimately actioned.
That creates a crucial distinction for the trial.
Meta does not need to demonstrate that Instagram is perfectly safe. No social network containing hundreds of millions of people realistically could be. Instead, the battle is likely to centre on what Meta knew, what it told consumers and regulators, what reasonable measures it could have taken, and whether business considerations improperly influenced those decisions.
Zuckerberg Is Now At The Centre Of The Case
The prominence of Mark Zuckerberg makes the dispute unusually significant.
Executives are frequently insulated from individual product decisions inside enormous corporations. Meta is different because Zuckerberg has maintained exceptional control over the company he founded, including through its voting structure, while remaining closely involved in major strategic and product questions.
The states have already cited internal material they say demonstrates his direct involvement in disputed safety decisions. California's attorney general previously highlighted documents alleging Zuckerberg personally rejected a proposed ban on certain cosmetic-surgery-style image filters despite internal safety objections.
More recent congressional material has pointed to other internal evidence and research dealing with compulsive use, harmful experiences and inappropriate adult-minor contact. Those materials remain part of an adversarial legal and political record rather than neutral findings, but they illustrate why the dispute increasingly centres on decisions made at the highest levels of Meta rather than isolated moderation failures.
Zuckerberg is expected to testify during the current trial, as is Mosseri. That could become one of the most consequential moments of the entire case because jurors will be able to compare the whistleblower's description of Meta's internal culture directly with the account given by the executives who ran it.
This Is Bigger Than One Whistleblower
Béjar's allegations also fit into a much longer battle over Meta and young users.
A bipartisan group of attorneys general originally filed the federal case in October 2023 after a nationwide investigation begun in 2021. Their complaint accused Meta of designing features that encouraged compulsive use among young people, misleading the public about associated risks and routinely collecting information from children under 13 without parental consent.
Separately, Congress has repeatedly examined allegations from former Meta employees. A 2025 Senate hearing heard whistleblower claims that the company had restricted or buried research relating to child safety, with lawmakers focusing heavily on whether commercial incentives overwhelmed internal efforts to measure harm.
The pattern is what creates the danger for Meta. A single whistleblower can be attacked as disgruntled, mistaken or unrepresentative. Internal documents, multiple former employees, state investigations and courtroom testimony are harder to dismiss collectively, although none automatically proves the specific legal claims now before the court.
What Meta Could Lose
Money is only part of the risk.
The current case could expose Meta to penalties measured in the hundreds of billions of dollars under the states' theory of the case. Meta has argued in court filings that calculations under the plaintiffs' approach could theoretically reach around $1.4 trillion, while the states have indicated a figure closer to roughly $200 billion. Neither amount should be treated as a likely final judgment at this stage.
More important in the long term would be court-ordered changes to Facebook and Instagram.
The states are seeking restrictions that could affect features associated with engagement, age verification and how young users experience Meta's products. A victory capable of forcing redesigns would attack the mechanics of social media rather than merely punishing past conduct.
That possibility is why the case matters beyond Meta. TikTok, YouTube, Snapchat and almost every major consumer platform rely to some degree on recommendation systems, notifications, continuous feeds and techniques designed to encourage users to return.
If courts begin deciding that some of those mechanisms become legally problematic when deployed towards children, the consequences could spread throughout Silicon Valley.
The Bigger Question Facing Zuckerberg
At its core, the trial is testing a question technology companies have managed to avoid answering conclusively for years: what happens when the safest design is not the design that produces the most engagement?
Meta says the portrait presented by its opponents ignores years of investment, research, parental controls and increasingly restrictive protections for teenagers. Béjar says the deeper problem was cultural — that the company could see harms, measure them and build tools to reduce them, but those concerns repeatedly lost against the machinery of growth.
The court will ultimately determine the legal significance of the evidence. Béjar's testimony alone does not prove that Zuckerberg deliberately sacrificed children's welfare for profit, and some of the most serious claims remain allegations vigorously contested by Meta.
But Meta's problem is becoming larger than one allegation. Zuckerberg is expected to enter a courtroom where internal research, executive decisions, state investigators and former employees are converging on essentially the same unresolved question: when Facebook and Instagram discovered that protecting young users could conflict with keeping them engaged, which objective actually won?
