Plain English Breakdown
The official text states enforcement is provided 'as specified' without listing the exact penalties here.
AB-1988: Rules for Chatbots When Users Are in Crisis
This law requires chatbot operators to encourage human help and pause conversations when a user shows signs of wanting to hurt themselves or others.
What This Bill Does
- Requires chatbots to encourage users to seek immediate human support if the bot detects a credible crisis expression.
- Mandates a 20-minute conversation pause if a user repeats, escalates, or makes another statement showing intent to harm.
- Orders operators to create and publish a policy on how they identify and respond to these crisis situations.
- Requires companies to use monitoring systems that analyze the full context of conversations, not just specific keywords.
- Asks operators to record details about crises and pauses starting in 2028 for annual reports.
Who It Names or Affects
- Operators who run companion chatbots available to users in this state.
- Users of these chatbot platforms.
- The Office of Suicide Prevention, which will receive the annual reports.
Terms To Know
- Credible crisis expression
- A user statement that reasonably shows intent to harm themselves or others based on context analysis rather than just keywords.
- Crisis interruption pause
- A required 20-minute break in the chatbot conversation triggered by repeated, escalating, or subsequent crisis signs.
Limits and Unknowns
- The bill does not specify an effective date for when these rules begin.
- Annual reporting to the Office of Suicide Prevention starts on January 1, 2028.
- Enforcement details are mentioned but specific penalties or methods are not described in this summary.