An agreement worth up to $18 billion ends the dispute between Meta and nearly all US states. Meanwhile, protecting children risks further expanding online data collection.
On 26 August 2026, Meta reached an agreement worth up to $18 billion with nearly all US states that accused it of designing Facebook and Instagram to increase children’s engagement, publicly downplaying the risks, and collecting data from children under 13 without parental consent.
The agreement ends the federal trial that began on 18 August and introduces new restrictions for teenagers, including usage limits and more effective systems for determining their age.
The case began in 2023, when 33 attorneys general sued Meta in federal court. It centred on two issues: the design of Instagram and Facebook – recommendation algorithms, infinite scrolling, notifications and other features accused of encouraging compulsive use – and the alleged violation of the federal law protecting the data of children under 13.
A significant part of the allegations emerged from the context revealed in the 2021 Facebook Files. Frances Haugen, a former Facebook product manager, provided thousands of internal documents showing that the company had conducted studies on the effects of its products, including Instagram’s impact on some teenagers. Political pressure increased at the time, Haugen testified before Congress, and Facebook suspended the Instagram Kids project intended for under-13s.
Facebook’s problems with data management, however, predate the Facebook Files by many years. The Cambridge Analytica scandal broke in 2018: an app collected data from millions of Facebook profiles, including people who had not used the app directly, and the British company later used it for political profiling activities. The episode is also covered in the 2019 Netflix documentary The Great Hack.
In 2019, the Federal Trade Commission (FTC) fined Facebook $5 billion, then the largest penalty ever imposed for privacy violations. According to the agency, Facebook had misled users about how it handled their data and violated an earlier privacy agreement with the FTC from 2012.
The trial concluded this week therefore concerns specific allegations involving Meta, children, platform design and personal data, but follows more than 15 years of disputes over how Facebook, and later Meta, collects, uses and protects its users’ information.
One of the clearest consequences of this debate has been the introduction of increasingly strict age limits for access to social networks. Australia was the first country to take the most radical step: from 10 December 2025, platforms including Instagram, Facebook, TikTok, YouTube, X, Snapchat and Reddit must prevent under-16s from holding an account and adopt systems to verify their age. In the following months, several countries moved in the same direction. Greece, Spain, France, Denmark, New Zealand and Slovakia are moving towards similar restrictions.
The contradiction is clear: to protect children from platforms accused of improperly collecting their data, we are building systems that require even more data from the people who use them.
How is age verified? What data is required? Where is it stored? For how long? And above all, can it be used for other purposes? The answers to these questions appear to interest no one and, above all, remain unavailable.
Following this case also reveals another paradox.
If the issue brought before the court also concerns how the product was designed – infinite scrolling, autoplay, push notifications, engagement metrics – why should one of the main responses be to identify users more effectively rather than change the product and create a better one?
A social network built to maximise the time spent on it and profile users in order to monetise their attention is a design and business-model choice, not an inevitable feature of technology.
Why is the Meta case not opening a broader debate about the social-network model itself? Is this really the internet we want to leave to the generations we say we are protecting today?
And above all, why is no one asking whether anyone is building better alternatives?
