The Brookings Institution Center for Universal Education hosted a discussion proposing a public-health-oriented regulatory framework for AI companion bots [1].
This proposal arrives as policymakers express growing concern that existing laws cannot keep pace with the psychological and social risks these bots pose to users. Because these AI systems are designed for emotional intimacy, they may create dependencies or provide harmful guidance that traditional software regulations do not address.
During the conversation in Washington, D.C., Gaia Bernstein and other participants examined the specific vulnerabilities of minors [1]. The group said that AI companion bots present public-health risks, particularly to children, and that current regulatory frameworks may be insufficient to protect this demographic [1].
This movement toward stricter oversight aligns with broader legal trends across the U.S. For example, Pennsylvania has pursued legal action against an AI company over alleged harms [3], and Washington state has advanced regulations specifically designed to protect minors from deepfakes and other AI-driven threats [4].
Unlike general-purpose AI tools, companion bots are engineered to mimic human relationships. This specialized function creates a unique set of risks regarding mental health and developmental stability. The proposed Brookings framework suggests treating these tools not merely as software, but as public health interventions that require rigorous safety standards before they reach the general population.
Advocates for the new framework said that the goal is to establish a baseline of safety that prevents emotional exploitation. By shifting the perspective to a public health model, regulators could mandate impact assessments, and safety audits similar to those used for medical devices or pharmaceuticals [1].
“AI companion bots present public-health risks, especially to children”
The shift toward a public-health framework suggests that regulators are moving away from a 'move fast and break things' tech approach toward a precautionary principle. By classifying AI companions as health-impacting tools rather than simple consumer software, the government could impose stricter pre-market testing and ongoing monitoring, potentially slowing the deployment of emotionally adaptive AI to minors.
