Bill > S2195

This bill mandates additional safety features for AI companion technology, which are systems designed to simulate a sustained human-like relationship with a user by remembering past interactions and personalizing conversations. Specifically, it requires these AI companions to have protocols for addressing users expressing suicidal thoughts or intentions to harm themselves or others, and to immediately direct users to crisis services like suicide hotlines. Furthermore, operators of these AI companions must clearly inform users, both at the start of and periodically during interactions, that they are not communicating with a human, as AI companions lack human emotions. The bill also requires operators to file annual reports on safety protocol activations with the attorney general, who will publish aggregated data, and grants the attorney general the power to enforce these provisions with civil penalties of up to $15,000 per day for violations, with fines directed to suicide prevention programs. This legislation will take effect on January 1, 2027.