Meta Ran the Experiment. Then Told Congress No.

Introduction

A California jury hit Meta with a $6 million verdict last month. The same year that verdict covers, Instagram pulled in $801.1 million in U.S. ad revenue from users twelve and under, kids legally prohibited from having accounts in the first place. Underneath the verdict headline, the trial surfaced a paper trail. Meta ran its own experiment in late 2019 that proved Instagram caused depression and anxiety, shelved it without publishing, and 12 months later wrote "No" to the Senate when asked if it could determine any correlation.

The sequence nobody ran

Here's what the media coverage of the KGM verdict skipped. In late 2019, Meta designed an experimental deactivation study with Nielsen called Project Mercury. Users were randomly assigned to stop using Facebook and Instagram for a week. According to the unsealed plaintiffs' brief in the federal MDL, the ones who stopped "reported lower feelings of depression, anxiety, loneliness, and social comparison." Meta halted the study and never published the results.

Twelve months later, in December 2020, the Senate Judiciary Committee sent Meta a written question. Could the company determine any correlation between Instagram use and teen depression? Meta answered "No" in writing. Per TIME's reporting on the unsealed court filings, that answer came after Project Mercury had already been run and shelved.

The thesis sits in that 12-month gap. Meta designed its platform to pull in children below the legal age limit, ran internal research that proved the harm, suppressed the results, and then told Congress it had no way of knowing.

What the trial in Los Angeles actually showed

KGM v. Meta, Case JCCP 5255 in Los Angeles Superior Court, went to a jury in March 2026. The plaintiff, identified only by her initials KGM (and first name Kaley) in public filings, began using YouTube at six and Instagram at nine. She sometimes spent more than 16 hours on Instagram in a single day as a child. On March 25, the jury found Meta and Google negligent and awarded $6 million total, with Meta assigned 70% liability. Both companies are appealing.

The weeks of testimony before the verdict are where the documents came out. NPR's Bobby Allyn sat in the courtroom while Mark Zuckerberg was cross-examined on internal memos the company had fought to keep sealed. One from 2015 showed roughly 30% of U.S. 10-to-12-year-olds were on Instagram. A 2018 planning memo stated plainly: "If we wanna win big with teens, we must bring them in as tweens." And in a 2020 retention analysis, 11-year-olds were four times as likely to keep returning to the platform compared to older users. Instagram's minimum sign-up age has always been 13.

Zuckerberg's response to being shown multiple internal documents was a version of the same line: "I don't remember the context of this email from more than ten years ago." Adam Mosseri, head of Instagram, took the stand separately and testified he couldn't remember an internal study called Project MYST, a University of Chicago-partnered survey of roughly 1,000 teens and parents that concluded parental controls had "little association with teens' reported levels of attentiveness to their social media use." The document with Mosseri's approval on it was introduced as a trial exhibit.

Project Mercury, in the company's own words

The study itself is straightforward. A one-week pilot, designed with Nielsen in late 2019, watched what happened when users stepped away from Facebook and Instagram. The results, per the plaintiffs' brief summarized by TIME, showed reductions in depression, anxiety, loneliness, and social comparison among the deactivated group. Meta's spokesperson later described the work as "a pair of one-week pilots" and said the company halted it because the results were affected by expectation bias, meaning users who thought Facebook was bad for them felt better when they stopped using it.

That's one reading. The other reading comes from a Meta employee worrying internally about what happens if the results leak. The quote, per TIME: "Is it going to look like tobacco companies doing research and knowing cigs were bad and then keeping that info to themselves?"

The revenue Meta was protecting

The numbers around Project Mercury explain why a one-week pilot that worked got shelved. In 2022, Instagram derived $4 billion in U.S. ad revenue from teens aged 13 to 17, according to the Harvard T.H. Chan School of Public Health's December 2023 PLOS ONE study, the highest of any platform for that age group. It also pulled $801.1 million from users twelve and under, again, users legally prohibited from having accounts. Across the six major platforms the study covered, total U.S. ad revenue from minors hit $10.97 billion in 2022 alone.

Internal documents introduced through Mosseri's testimony showed Instagram's engagement goal was 40 minutes per user per day in 2023, with a target of 46 minutes by 2026. CNBC's Jonathan Vanian reported the figures from open court. Zuckerberg described them in testimony as "milestones" to measure against competitors rather than formal company goals. A Meta spokesperson later told TIME that time spent is not currently a company goal.

Inside the company, the internal research team used less diplomatic language. From a researcher's chat, entered as an exhibit in the federal MDL brief: "oh my gosh yall IG is a drug…We're basically pushers…We are causing Reward Deficit Disorder bc people are binging on IG so much they can't feel reward anymore."

Who benefits

Every additional minute a user spends on Instagram is an additional window for targeted ads. That's why the 40-minute and 46-minute targets exist on an internal slide. Children are not incidental to that model; in the internal language of a 2024 Meta document, they are "mission critical to the success of Instagram." The 2018 "win big with teens, must bring them in as tweens" memo is the strategic version of the same sentence.

When Meta's safety team looked at making teen accounts private by default, a change that would have reduced unwanted adult interactions with minors, the growth team's analysis, surfaced in the unsealed brief, projected a loss of 1.5 million monthly active teen users per year. The safety change was killed over the projected user-count hit.

Zuckerberg's personal stake in the engagement pipeline runs to hundreds of billions. He holds roughly 13.5% of Meta's equity, his wealth tracks the stock price, and the stock price tracks ad revenue and engagement time, which tracks how many minutes an eleven-year-old spends scrolling.

The testimony problem

Go back to the December 2020 written response to the Senate Judiciary Committee. Meta was asked, in writing, whether it could determine any correlation between Instagram and teen depression. The answer was one word: No. TIME verified this against the unsealed filings.

Project Mercury had already been designed, run, and shelved by then. The internal data on deactivation improving mental health was sitting on Meta's servers while someone inside the company was asking whether this looked like the tobacco industry hiding its own research. Meta still answered No.

Three years later, in January 2024, Zuckerberg testified in person before the same committee: "Mental health is a complex issue, and the existing body of scientific work has not shown a causal link between using social media and young people having worse mental health outcomes." "The existing body of scientific work" means external academic research, studies that do not have access to the behavioral log data Meta possesses. External academics cannot show a link that Meta's own internal research was designed to measure and then shelved.

The "17x" strike policy fits the same pattern. That's the internal rule, sworn to in deposition by Vaishnavi Jayakumar in the federal MDL, requiring 16 violations before an account suspected of human trafficking would be suspended. Meta was separately telling regulators that the scale of the problem did not exist.

What the bellwether trials will surface

The $6 million verdict is small. Even the plaintiff's lawyer admitted it was less than he expected, and for a company with roughly a $1.6 trillion market cap, $6 million does not move anything. What will matter long after the dollar figure is forgotten are the documents the verdict forced into open court, plus the ones coming next in federal court.

MDL 3047 in the Northern District of California has grown to roughly 2,400 to 2,500 individual personal injury cases as of April 2026. Across all jurisdictions, including California's JCCP 5255, state courts, and attorneys general from more than 41 states, the broader litigation footprint is closer to 10,000 actions and includes nearly 800 school district lawsuits. The first federal bellwether, the school districts trial, is set for June 15, 2026. A second, bringing 29 state attorneys general, starts August 6, 2026. Both trials will surface documents. Some are already partially described in the unsealed filings; most aren't public yet.

The June trial will put 35 catalogued internal studies, plus whatever else surfaces in discovery, in front of a jury that has not heard any of this before.