Inside the Files Meta Buried for Years & How Zuckerberg Explained Them to Court

On February 18, Meta’s CEO Mark Zuckerberg faced a jury for the first time over allegations that his platforms were deliberately designed to addict children. A trove of damning internal documents, some revealed publicly for the first time, is laying bare what the company knew and when.

Their predator funnel, tolerance to sex trafficking, the under-13 problem, and impact on body image and mental health…

Zuckerberg was grilled for hours about whether the company he built deliberately hooked children on its platforms and then hid the evidence. Outside the courthouse, parents who say they lost children to social media-related tragedies held photographs of their kids. Inside, a plaintiff’s lawyer unspooled a 35-foot-wide collage of selfies posted to Instagram by the young woman suing Meta, a visual rebuke to years of Zuckerberg’s congressional testimony that his platforms exist to foster “meaningful social connections.”

The case is a bellwether trial tied to more than 1,600 similar lawsuits brought by families and school districts across the United States. Its outcome could shape not just Meta’s legal exposure, potentially in the billions of dollars, but the future of how social media platforms are designed, regulated, and held accountable. TikTok and Snapchat have already settled; Meta and Google’s YouTube remain as defendants, with the trial expected to run six to eight weeks.

What makes this trial different from years of congressional testimony and regulatory scrutiny is what’s now on the public record: internal documents, chat logs, and design memos that the company fought for years to keep sealed. Taken together, they sketch a portrait of a corporation that knew its products were harming children, studied the damage in detail, and in key instances chose profit over intervention.

What the Company’s Own Employees Said

Among the most striking documents now in evidence are internal Meta chat logs in which a UX researcher wrote that Instagram is “a drug” and that employees were “basically pushers.” The quote is a single employee’s offhand remark, but it sits alongside a broader body of internal research that the company commissioned, buried, and in some cases actively suppressed.

Plaintiffs’ attorneys allege that Meta’s engineers did not accidentally create high engagement — they engineered it systematically, deploying what psychologists call variable reward schedules: the same mechanism used in slot machines, where unpredictable payoffs drive compulsive behavior. 

Features like infinite scroll and algorithmically personalized content feeds are, according to expert witnesses cited in court filings, specifically calibrated to override the developing impulse-control systems of adolescent brains.

Meta disputes this characterization. During his testimony Wednesday, Zuckerberg argued that greater engagement with Instagram reflects the platform’s value to users, not a manipulation of their psychology. “There’s a misconception that the more attention the company captures, the better it is for Meta’s bottom line, regardless of harms,” he said. Plaintiff’s counsel responded by presenting a 2023 internal goal to push average daily user engagement to 40 minutes, and a 2026 target of 46 minutes.

The Under-13 Problem: Four Million Children

Instagram’s official minimum age is 13. The multistate attorney general complaint alleges that by 2015, Meta’s own internal analysis found roughly 4 million children under that threshold were already active on the platform, a figure representing approximately 30 percent of that age group in the United States. 

On Wednesday, lawyers displayed that document in court; Zuckerberg acknowledged that some users misrepresent their age during sign-up, and said the company removes underage users it identifies. The plaintiff’s lawyer’s response was pointed: “You expect a 9-year-old to read all of the fine print?”

Court filings further allege that some within the company viewed the 10-to-12-year-old demographic as commercially valuable, and that one internal memo cited company strategy in terms similar to those used by tobacco executives. 

“Seriously saying ‘we have to hook them young’ here”, from inside company chats. 

These are allegations from plaintiffs’ filings, which Meta contests. But they are now part of a public record a jury is weighing.

Body Image Research: What Meta Knew

The internal research on teen girls and body image is among the best-documented elements of this litigation, in part because it was first reported by the Wall Street Journal in September 2021 based on documents provided by whistleblower Frances Haugen. 

Meta’s own slides reportedly showed that roughly one in three teen girls said that when they felt bad about their bodies, Instagram made them feel worse. 

Separately, according to the multistate AG complaint, internal research found that certain “beautification” filters were associated with body dysmorphia in young users.

What is alleged to be new — and what the current trial is in part designed to test — is not the existence of this research but what Meta chose to do with it. Plaintiffs allege that filters linked to body image harm were retained because they drove higher engagement, a claim Meta disputes. 

The company’s position is that it has introduced extensive safeguards for younger users in recent years and that the causal link between Instagram use and mental health outcomes has not been scientifically established.

How the Algorithm Routed Children to Groomers

Among the most explosive allegations now before the court (and among the least covered in earlier rounds of reporting) is what plaintiffs describe as Meta’s “predator funnel”: the company’s own recommendation algorithm actively connecting minors with adults flagged internally as potential child predators. The mechanism was not a bug. According to plaintiffs, it was the product working as designed.

The figures come from the plaintiffs’ corrected omnibus opposition brief filed in the Northern District of California, and are drawn from Meta’s own internal data. The company’s internal shorthand for these encounters was “IIC” — “inappropriate interactions with children” — a bureaucratic acronym that, according to the brief, became common enough to acquire its own abbreviation inside the company.

Meta’s own safety researchers reportedly recommended as early as 2019 that teen accounts default to private mode, which would have substantially curtailed unwanted contact from strangers. According to court filings, the growth team rejected the proposal because it would “likely smash engagement.” The default private setting for all teen accounts was not implemented until 2024, five years later. In the intervening period, plaintiffs allege, teens experienced billions of unwanted interactions with adult strangers.

The 17-Strike Rule: Meta’s Tolerance for Sex Trafficking

If the predator funnel allegation concerns what the algorithm did, the “17-strike” policy concerns what the company chose not to do about it. According to sworn testimony from Vaishnavi Jayakumar, Instagram’s former head of safety and well-being, Meta operated under an internal policy that allowed accounts engaged in the trafficking of humans for sex to incur 16 violations before being suspended on the 17th. Jayakumar testified she learned of the policy when she joined the company in 2020 and was immediately alarmed.

“You could incur 16 violations for prostitution and sexual solicitation, and upon the 17th violation, your account would be suspended,” Jayakumar testified, adding that the threshold was “by any measure across the industry, a very, very high strike threshold.” 

Plaintiffs say her testimony is corroborated by internal company documentation.

The policy becomes more striking in context: at the same time, Instagram allowed users to easily report comparatively minor violations: spam, intellectual property violations, promotion of firearms, directly within the app. Reporting child sexual abuse material, by contrast, offered no simple in-app mechanism, an issue Jayakumar says she raised repeatedly and was told would be too difficult to fix.

Meta’s response: the company told USA Today it now operates a “one-strike” policy and removes accounts immediately upon determination of the most severe trafficking violations. The company says the strike threshold was lowered progressively over time after 2019. It did not dispute that the 17-strike policy existed.

Eight People for Eight Billion Users: The Staffing Reality

The predator funnel and the 17-strike policy share a common backdrop: a child safety operation that, by plaintiffs’ account, was catastrophically under-resourced relative to the scale of the problem. According to court filings, at one point Meta assigned just eight people — five engineers, two data scientists, and one data engineer — to cover all child safety problems globally across its platforms, which at the time had billions of users.

A separate internal document cited in the plaintiffs’ omnibus opposition describes child safety as “explicitly a non-goal” for a particular product team. 

Adam Mosseri, the head of Instagram, testified last week that he does not believe people can be clinically addicted to social media. “I think it’s possible to use Instagram more than you feel good about,” he said. “Too much is relative — it’s personal.” That framing — locating responsibility with the user rather than the platform — is at the heart of Meta’s defense.

The Legal Stakes: A Pivot Point for the Industry

Meta is fighting this battle on multiple fronts. While the Los Angeles trial plays out, eighteen state attorneys general are separately challenging the company in the Ninth Circuit Court of Appeals. In January, the appeals court judges signaled skepticism toward Meta’s core legal shield — a federal law called Section 230 that tech companies have long used to avoid liability. The judges appeared unlikely to dismiss the case, meaning more legal exposure ahead.

Lawyers have compared this case to the tobacco lawsuits of the 1990s, where internal documents eventually made denial impossible. This trial is the first real test for thousands of similar cases waiting in the pipeline. If the jury sides with plaintiffs, pressure to settle across all of them grows fast.

Meta and Google are pushing back. Meta’s lawyers argue the plaintiff had mental health problems before she ever opened Instagram, and that the app actually helped her cope. Google, meanwhile, is arguing that YouTube isn’t even a social media platform — a technical claim that, if a judge accepts it, could get them off the hook entirely.

The verdict will directly affect 1,600 other cases — and could force billion-dollar payouts or fundamental redesigns of how these platforms work. What’s clear from Meta’s own documents, its own data, and the testimony of its own former employees is this: the company made choices. Now a jury gets to judge them.

Sources

1. Opening-Answer Brief of the State Attorneys General (Case No. 24-7032)

2. Multistate Attorney General Complaint against Meta Platforms, Inc.

3. Internal Meta Chat Logs and Exhibit 74 (MDL No. 3047)

4. Plaintiffs’ Corrected Omnibus Opposition to Defendants’ Motions for Summary Judgment

5. California Superior Court Minute Order (Case No. 22STCV21355)