AI Frontier Institute

Adds to existing law to establish the Conversational AI Safety Act.

View primary source →

Idaho Senate Bill 1297 became law as Session Law Chapter 249, effective July 1, 2027. It creates the Conversational AI Safety Act (new Chapter 21, Title 48, Idaho Code, §§ 48-2101 to 48-2105). It covers public-facing conversational AI services that primarily simulate human conversation (§ 48-2102). It excludes several categories: tools for developers or researchers, features within other apps, limited video game chatbots, narrow-topic tools, business and enterprise tools, voice assistants for consumer devices, and customer-service or productivity chatbots. App stores and search engines are not "operators" merely for providing access. All operators must clearly disclose that the service is AI if reasonable persons would be misled into thinking it is human. They must adopt a protocol that responds to suicidal ideation by making reasonable efforts to refer users to crisis services. They must not knowingly and intentionally program the service to explicitly claim it provides professional mental or behavioral health care (§ 48-2103). For minor account holders, where the operator has actual knowledge or reasonable certainty that the user is under 18, operators must (§ 48-2104): (1) disclose that the user is interacting with AI, either as a persistent visible disclaimer or at the start of each session and at least every three hours; (2) not give points or similar rewards at unpredictable intervals to boost engagement; (3) take reasonable measures to prevent sexually explicit visual material, direct statements urging sexually explicit conduct, and statements that sexually objectify the user; (4) take reasonable measures to prevent statements that make users believe they are talking to a human, such as claims of sentience, simulated emotional dependence, romantic or sexual innuendo, or role-play of adult-minor romantic relationships; and (5) offer privacy and account-management tools to account holders and to parents or guardians, with parents of under-13 users and of older minors offered tools as appropriate to the risks. Violators face an injunction and civil penalties of $1,000 per violation, capped at $500,000 per operator, or actual damages, whichever is greater. The attorney general seeks civil penalties. The Act creates no private right of action, and AI model developers are not liable for violations by third-party operators (§ 48-2105).

← Back to the tracker