US jury sees Meta chat over in-house lawyers' purge of teen safety documents

August 25, 2026, 22:50 GMT | Insight Instagram's algorithms sent content to teen users about eating disorders, suicide, bullying, hate, violence and nudity at a much higher rate in recent years than to adults, even as company lawyers scrubbed those previously unreported statistics from company documents to avoid legal scrutiny, a federal jury in California heard Tuesday.

Instagram head Adam Mosseri, the senior-most Meta Platforms executive to take the witness stand so far in the company's trial against 29 state attorneys general over addictive design and privacy claims, pushed back aggressively on that narrative, denying that the social media giant intentionally withheld from regulators and other plaintiffs suing the company key information about the allegedly addictive design of Facebook and Instagram, and the low adoption rate of its safety tools.

"This is all news to me," Mosseri said when confronted with a 2023 internal chat, where employees working on teen safety issues at Instagram reacted with apparent consternation as a growing scrum of lawyers began removing data such as teens on Instagram having a 1.5 times higher rate of seeing content about bullying, suicide and self-harm, hate, and nudity, and 2.5 times the rate of eating-disorder content than adults — data that a lawyer for the state had not made public before Tuesday.

"It does not surprise me that people are careful about information that comes before my desk," Mosseri said, but he asserted he has never "encouraged people to bring less information to me," and that the goal of legal review was to make sure he and other top executives had the best possible information to map company strategy.

"In general for any complicated issue, we want to make sure that all of the experts can weigh in, and there are often legal implications of the product work we do, particularly around compliance," Mosseri said.

In a striking real-time battle over evidence issues that are typically decided months before a trial, the states only got access to the full record of the July 7, 2023, chat during the course of Tuesday's trial proceedings, as US District Judge Yvonne Gonzalez Rogers overruled earlier discovery rulings by a federal magistrate judge that parts of the chat were protected from disclosure by attorney-client privilege.

Gonzalez Rogers said the redactions were improperly made for business advice provided by the lawyers, instead of for legal advice, a ruling Meta lawyer Ashley Simonsen protested vigorously.

— The 2023 chat —

Throughout the proceedings, the states showed the advisory jury and Gonzalez Rogers, the ultimate fact-finder in the trial, a series of internal documents that go to the heart of the states' claim that Meta deceptively told users its products were safe when leadership knew the addictive design of Facebook and Instagram were harming the mental health of millions of young users.

Many of the internal documents, including the chat in 2023 — which occurred at a time when the states' investigation of Meta was already advanced, just a few months before the states filed suit — have not been used as evidence in previous trials against Meta this year in Los Angeles, New Mexico and Tennessee.

In the chat, between Francesco Fogu, director of product design at Instagram, and other Instagram teen well-being members, the group appeared to be upset about the extensive editing by the company's lawyers, removing language such as "Teens and parents are worried about social media," and replacing it with the language that they "care deeply" about social media.

"Shoot, she's tearing apart some of our strongest slides," one Instagram well-being team member complained as one company lawyer, soon joined by at least five other lawyers, began editing a slide deck that the team was preparing to present to Mosseri and Max Eulenstein, head of product for Instagram.

"I mean these are pretty strong [lawyer changes] and could change the outcome of the discussions. Does legal not want us to share them at all?" Fogu wrote in the chat.

"If I had to guess, they don't want a paper trail tracing this back to information visually presented to leadership that may have to testify," wrote Meagan Palatino, another member of the well-being team.

"Apparently, we need to limit their exposure to this information," Palatino said later in the chat, copying instructions she received from an in-house lawyer for the group that: "For Max or higher, we should try to get it buttoned up so there's no litigation exposure risk for Max or Adam on what is in this deck."

"This is just of course to avoid more facebook files," another member of the Instagram well-being team, Cristina Foung, wrote in the typo-laden chat, referring to the Wall Street Journal's investigative series in 2021 following the disclosures of Facebook whistleblower Frances Haugen.

Lawyers for the states showed the text of the chat conversation to the jury and Gonzalez Rogers through the live testimony of Fogu, an 11-year veteran of Meta who consistently resisted the states' line of questioning and disputed that the lawyers' replacement of "worried" with "caring deeply" was a significant change.

"I think it's the same thing," Fogu told Jason Slothouber, a lawyer for the Colorado Attorney General.

"'Caring deeply' could express a whole set of emotions," Slothouber countered, in one of the many antagonistic exchanges between the two.

"I think they are related," Fogu said.

Slothouber asked Fogu why Meta hadn't disclosed the higher rates of harmful content being shown to teens compared to adults. "To the best of your knowledge those statistics weren't public until you sat in that chair" and testified to them, Slothouber said.

"I don't know, sir. They could have been public through other means," the Meta executive said.

Fogu was so resistant to some of Slothouber's examination that Gonzalez Rogers was drawn into directly questioning him in concert with the Colorado lawyer at one point, expressing vocal skepticism when the Instagram official said he didn't know facts such as whether the initially low adoption rate of safety tools had increased.

"You're the director of this and you don't know if this is the highest this has ever been?" the judge asked Fogu incredulously about trends in the adoption rate of Instagram safety tools.

"Mmm, interesting," Gonzalez Rogers remarked after Fogu responded "I don't know" multiple times in a row to her and Slothouber's questions. Members of the jury could be seen scribbling notes industriously during the exchange.

— Other Meta documents —

The contents of the July 2023 chat were just one in a series of Meta internal documents lawyers for the states showed to the jury in which members of the company acknowledged that safety products.

"Our current strategy primarily uses opt-in based adoption and low friction due to constraints to minimize harm to the ecosystem. This has been ineffective at both creating a strong drumbeat externally and reducing PU behaviors," one Meta document from 2023 seen by the jury said, with "PU" referring to "problematic use." 

Another document in which Fogu was listed as lead author, but which he testified that he couldn't remember writing, said Meta doesn't always comply with regulatory orders.

"As a company, we make different choices when it comes to regulatory responses," the document said. "Sometimes we don't comply and accept a fine. Sometimes we comply in the most minimal or literal way possible."

In another document the jury saw Tuesday, Meta said its "current strategy" of opt-in status had kept a "modest" awareness by teens of its safety tools, as a way "to minimize ecosystem impact."

The strategy, the document said, was to keep "a high bar on which users were targeted and how often, with long cooldowns [about safety tool notifications]. As a result, surveys show awareness of our products is modest — only 33 percent of teens are aware of 'Take a Break' and 44 percent of 'Quiet Time.'"

— Meta documents —

Meta lawyer Moira Penza countered during her cross-examination of Fogu that there were other internal documents from 2023 showing how committed the company was to teen safety.

"Across IGWB and IG Trust, we've been working on a refreshed vision for teen protection on IG, looking beyond youth industry practices and near-term regulatory requirements to define which areas we should lead the industry," the Instagram well-being team said in an email to Mosseri and Eulenstein in 2023. The acronyms referred to "Instagram Wellbeing."

Fogu said that while the lawyers did make significant changes referenced in the 2023 chat, even with all the data they removed, the key result was that the meeting with Mosseri led to Meta launching a bundle of protections called Instagram Teen Accounts in 2024.

"If it was removed from the beginning of the deck, we made sure to have it later in the deck," Fogu said of the chat redactions.

As a result of the meetings with Mosseri, "the team was able to push through the creation of Teen Accounts," Fogu said, "an even better solution for protecting teens."

Please email editors@mlex.com to contact the editorial staff regarding this story, or to submit the names of lawyers and advisers.