Task force endorses guardrails for AI 'companion' chatbots, emphasizes minors' protections

Crystal Leatherman, cochair of the task force’s consumer protection and public safety subcommittee, presented Recommendation 1, a set of proposed guardrails for generative AI companion chatbots intended for personal interaction. Leatherman said the recommendation emphasizes "human oversight," independent testing and "privacy safeguards for minors," and that much of the proposal aligns with legislation that recently passed into law.

Public commenters including Sachin Nanday told the task force the proposal is urgent for child safety. "From a citizen safety standpoint, recommendation 1 is urgent," Nanday said, urging strong enforcement and clear disclosures for minors using AI companions.

Task force members generally signaled support for the approach and praised the subcommittee’s stakeholder outreach. Members said the recommendations aim to be proportional to risk and narrowly targeted to companion chatbots — not to all chatbot or search‑engine interactions — to avoid overreach while protecting vulnerable users.

The task force did not vote on the recommendation at this meeting; members were invited to provide redlines and supporting evidence ahead of an April 24 vote. The subcommittee plans to incorporate public feedback and technical detail — for example, what kinds of independent testing and disclosure language should be required — before the final report is prepared for the legislature.