CHAT Act 2.0

Key Points

• The bill sorts companion AI chatbots into three tiers based on their purpose: educational tutors, companionship or friendship bots, and mental health or therapy-style bots, with different safety rules for each.
• Every chatbot used by a minor must clearly say it is not a human at the start of a conversation, and companionship bots must repeat that reminder every 60 minutes along with take-a-break notifications every 90 minutes.
• Parents get monitoring tools and must be told right away if their child expresses suicidal thoughts to a chatbot, and companies cannot use a child's emotional state or conversations to target them with ads.
• Therapy-style AI chatbots cannot be used by minors unless a licensed medical or mental health professional is supervising, and these bots are banned from giving diagnoses or acting as a stand-in for real mental health care.
• Minors get a simple way to delete their entire chat history with a companion bot, and companies generally cannot sell or share a minor's data with third parties without a parent's clear consent.