What happened
Brookings Institution scholar Gaia Bernstein published a May 2026 policy brief arguing that AI companion bots constitute a public health problem imposing three layers of harm: lack of guardrails, addictive design, and disruption of social skill development — particularly in children. The brief contends that the regulatory debate has been falsely framed around bans, which are both politically difficult to sustain and insufficiently targeted at products already on the market. Bernstein proposes reframing the legal approach around a public health model centred on pre-market approval and product recall authority, analogous to the Consumer Product Safety Commission's mandate for physical goods. The recall mechanism is presented as the most urgent practical lever because it can be applied retroactively to products already deployed, creating a de facto safety floor that incentivises manufacturers to conduct safety testing before launch.
Why it matters
As AI companion products proliferate — and as several high-profile harm events involving minors attract legislative attention — the framing Bernstein proposes is gaining traction with US state and federal policymakers; boards and legal teams at consumer AI companies should understand this regulatory trajectory before it reaches committee.
Action needed
Share with your regulatory affairs and product safety teams; assess whether existing companion-style AI features in your portfolio would meet a pre-market approval standard, and monitor US state legislative activity for recall-based proposals.