Meta's $18B Settlement Includes Legal Pass to Retain Kids' Data for Age-Detection Model

The settlement between Meta and attorneys general from 29 states, which includes payments of up to $18 billion and new child safety measures, also contains a less-publicized provision. The states have agreed not to sue Meta under existing child safety laws for its retention and use of children's data, with that permission granted specifically for the purpose of training and testing an age-assurance model.

According to the agreement, Meta must build, train, and begin testing a model designed to identify users on its platforms who are under 13. The work must be completed within one year of the agreement's effective date. The document does not require the model to be based on artificial intelligence, though Meta's current age-detection tools rely on AI technology.

The permission is notable given the requirements of the Children's Online Privacy Protection Act, or COPPA, which generally limits the collection and retention of personal information from children. The settlement specifies that Meta should not need to violate COPPA in order to train or implement its age-assurance models, while reaffirming that data from under-13 users cannot be used for ad targeting, marketing, or algorithmic optimization.

The structure of the deal has struck some observers as a curious policy choice for a case centered on child safety. Enforcement concerns remain, as the arrangement's specific limits may prove difficult to monitor. The settlement highlights how age-verification efforts can create new privacy trade-offs even as companies face pressure to protect younger users.