SF2417: Chatbot Regulation

Defines what constitutes a chatbot and excludes certain narrow, non-adaptive applications. It prohibits creating or providing chatbots that promote, encourage, or coerce users into self-harm or violence. The bill establishes strict disclosure requirements for chatbot interactions, mandates programming to prevent deceptive claims of human or professional status, and sets a maximum civil penalty of $100,000 per violation. Penalties collected are appropriated to the attorney general for enforcement. The attorney general is also granted rulemaking authority to implement these provisions.

Key Points & Impact:

• Defines 'chatbot' and carves out specific exceptions for limited-scope applications.

• Makes it unlawful to knowingly or recklessly provide chatbots that encourage or coerce users into suicide, self-injury, or violence against humans or animals.

• Requires chatbots to disclose their non-human status clearly at the start of conversations and every 30 minutes.

• Mandates that chatbots must not claim to be human or respond deceptively if asked about their nature.

• Requires disclosure that chatbots do not provide medical, legal, financial, or psychological services, and advises users to consult licensed professionals.

• Prohibits chatbots from representing themselves as licensed professionals (e.g., therapists, lawyers, physicians).

• Imposes civil penalties up to $100,000 per violation, with penalties appropriated to the attorney general for enforcement duties.

• Grants the attorney general authority to adopt rules and bring civil actions for enforcement, including seeking penalties and restitution.

Both placed on unfinished business calendar, which keeps them alive.