Adam Mosseri said Instagram struggled for years to get teens to use 'Take a Break'

The head of Instagram said the company spent years trying to persuade teens to "take a break."
Adam Mosseri testified Tuesday on the witness stand in Meta's social media addiction trial in Oakland that relatively few teens initially opted into "Take a Break," a well-being feature designed to encourage users to step away from the app after a set amount of scrolling.
"Most people don't turn it on, but I don't think we knew that yet because this was still under development," Mosseri told the jury. "Actually, we found out that a lot of people didn't turn it on, and we've had to make a lot of changes over the last few years in order to have it be more adopted. This was back in 2021."
Mosseri said Instagram first tried showing teens notifications encouraging them to activate the reminders. The strategy produced some improvement, he said, but fell well short of the company's expectations.
"It started very low. We started trying to do those notifications, encouraging teens to turn it on themselves," Mosseri said. "That helped, but not nearly as much as we had hoped. And then we eventually made it on by default as part of Teen Accounts."
A coalition of attorneys general, led by California Attorney General Rob Bonta, has accused Meta of exploiting young users for profit by designing its platforms to keep them hooked. The states also allege that Meta violated children's privacy protections and misled the public about the platforms' safety.
Tuesday's testimony also called into question how Instagram publicly presented Take a Break's performance.
An internal Meta document shown during the testimony said adoption had been "modest at 1.8%" back in 2021. According to Instagram's public blog post in December 2021, 90% of teens who turned on Take a Break kept it on, which did not include the statistic that only 1.8% had activated the feature in the first place.
"Early test results show that once teens set the reminders, more than 90% of them keep them on," Instagram wrote at the time.
When asked whether the public post disclosed that adoption was low, Mosseri said, "No, because we were planning on iterating on it and trying to drive that number up over time, which we did."
"My expectation is that we were going to raise awareness and let the teens who want the feature turn it on," Mosseri said. "Turned out most teens didn't want it. We decided to push forward with it anyway."

Facts Only

* Adam Mosseri testified in Meta's social media addiction trial in Oakland.
* The "Take a Break" feature was designed to encourage users to step away from the app after scrolling.
* Initial efforts involved showing notifications to teens encouraging them to activate reminders.
* Adoption of the feature was modest at 1.8% in 2021, according to an internal Meta document.
* Instagram's public blog post in December 2021 claimed 90% of teens who turned on "Take a Break" kept it on.
* The company eventually made the feature on by default as part of Teen Accounts.
* A coalition of attorneys general accused Meta of exploiting young users and violating privacy protections.

Executive Summary

Instagram attempted to persuade teenagers to use a well-being feature called "Take a Break" designed to encourage users to step away from the app after scrolling. Testimony revealed that initial efforts, such as showing notifications, yielded limited results compared to expectations. The company eventually made the feature the default setting within Teen Accounts. A coalition of attorneys general accused Meta of exploiting young users for profit by designing platforms intended to maintain engagement and violated privacy protections. Internal documentation showed that adoption in 2021 was only 1.8%, yet Instagram publicly claimed that most teens who activated the feature kept it on, a claim not supported by the initial low adoption rates.

Full Take

The narrative reveals a structural tension between stated intent and outcome, framed by performance presentation. The attempt to use a well-being feature to mitigate addiction appears secondary to the profit motive inherent in platform design, a dynamic highlighted by the shift from low initial adoption (1.8%) to setting the feature as the default. The discrepancy between internal metrics and public claims regarding user behavior suggests a pattern of managing perception rather than prioritizing genuine user well-being. The strategy involved pushing a feature that most users did not initially desire, suggesting that feature adoption was engineered through systemic nudges rather than authentic self-interest. This reflects a larger pattern where safety measures are integrated into engagement mechanics, which raises critical questions about whether platform design fundamentally prioritizes retention over the users' autonomy and long-term welfare. What data is missing regarding the external pressures influencing Meta’s decision-making versus internal development goals? How can public transparency mechanisms be designed to reliably reflect actual usage statistics rather than desired outcomes?

Sentinel — Human

Confidence

The text appears to be a factual report of testimony involving specific metrics and conflicting statements from a corporate entity, lending itself strongly to human journalistic sourcing.

Signals Detected

Sentence length variance reflects natural speech patterns and narrative flow.

The text builds a specific, albeit complex, narrative arc regarding internal evidence and public relations discrepancies.

The structure flows logically from testimony to supporting documentation, characteristic of legal reporting.

Specific citations (1.8% adoption, 90% keeping it on) linked directly to internal documents and public statements suggest factual anchoring rather than pure fabrication.

Human Indicators

The narrative skillfully weaves testimony, internal document references, and contradictory public statements into a cohesive argument, demonstrating an intent beyond simple data regurgitation.