Meta has reached a settlement agreement with attorneys general from 29 states, which includes a payment of up to $18 billion and the implementation of child safety measures. A notable aspect of the agreement is that the states have consented not to sue Meta under existing child safety laws regarding its retention and use of children's data. This provision is specifically for the purpose of training and testing Meta's age-assurance model, which is intended to identify users under the age of 13. The model must be developed and tested within a year from the effective date of the agreement. While the settlement does not require the model to be AI-based, Meta's current tools for age detection utilize AI technology.
Under the Children’s Online Privacy Protection Act (COPPA), websites and apps are generally required to limit the collection and retention of children's personal information. The settlement indicates that Meta should not need to violate COPPA to train or implement its age-assurance models. However, it also states that the state attorneys general have agreed to refrain from bringing any past, present, or future COPPA claims related to Meta's use of children's data.
The agreement specifies that Meta is prohibited from using data from users under age 13 for advertising, marketing, or algorithmic optimization. Philip N. Yannella, a partner at Blank Rome, noted that the request for legal protection from Meta and the states' agreement is not unreasonable, as data minimization practices are common in privacy compliance. However, he pointed out that COPPA is primarily enforced by the Federal Trade Commission (FTC), which is not a party to this settlement, raising questions about the FTC's stance on the matter.
The settlement requires Meta to isolate its understanding of children's behavior signals and use this data solely for the purpose of detecting and removing under-13 users. An independent auditor will monitor Meta's compliance with the settlement, ensuring that the company adheres to the terms. However, the enforcement of these limitations may be complex, as the data could potentially be integrated into other Meta systems over time.
The agreement's provision preventing state AGs from pursuing COPPA or similar claims in the future could complicate legal options if concerns arise regarding Meta's data usage. Joshua Wurtzel, a partner at Schlam Stone & Dolan LLP, stated that if Meta uses the data outside the agreed parameters, the release from liability would not apply, but such disputes could be intricate, depending on whether Meta's actions align with the settlement terms. Peter Jackson, a Data & IP attorney, expressed concern that this provision might discourage future enforcement actions.
The settlement also raises broader questions about data usage in the AI industry, as AI systems often require extensive access to personal data. Meta may need significant insights into children's social media usage to effectively identify accounts belonging to young users.