Meta’s $18 billion settlement with attorneys general from 29 states includes a controversial clause: the states agreed not to sue Meta for using children’s data to train its age-assurance model. The agreement allows Meta to collect and retain children’s data specifically for developing and testing the model, which must be operational within a year. The provision is meant to shield Meta from claims under COPPA and similar state laws, but it stops short of permitting ad targeting or algorithmic optimization using data from users under 13.
Legal experts note the settlement’s guardrails are standard for privacy compliance, though enforcement could prove difficult. Meta must isolate the data used for age detection, preventing it from feeding into other systems. An independent auditor will monitor compliance, but gaps remain: the agreement doesn’t specify what data Meta will retain, how much behavioral information it includes, or how long it will keep the data. Future changes to the model also remain unclear.
The decision raises broader questions about AI development and data access. Meta’s age-detection tools rely on deep insight into children’s social media use, but the settlement’s legal shield could complicate future enforcement if the company strays from the agreed terms.
