Senior Chatbot Protection Act of 2026

Key Points

• AI chatbots would have to clearly tell users they are talking to a machine, not a human, at the start of conversations and regularly during long chats. Companies could not let their AI pretend to be a licensed doctor, lawyer, or financial advisor.
• When a chatbot senses someone is making a major decision, like changing medical treatment or moving money, it must warn the user that it is not a licensed professional and that the advice should be checked with a real person first.
• If a chatbot detects signs of a mental health crisis or medical emergency, it must stop giving advice and instead direct the person to emergency services or a crisis line such as 988.
• Chatbot makers are banned from using manipulative design tricks that exploit memory loss, isolation, or other age-related vulnerabilities to keep older users hooked or discourage them from talking to real people.
• Users get the right to delete their chat history, and companies cannot use someone's conversations to train AI models without that person's clear, affirmative permission first.
• Companies that knowingly or recklessly break these rules can be fined up to $50,000 per violation, and both the FTC and state attorneys general can take them to court.