RELATING TO ARTIFICIAL INTELLIGENCE.
SB3001, the “Artificial Intelligence Disclosure and Safety Act,” creates a new set of consumer-protection rules for AI companions in Hawaii. The bill requires operators of AI companion systems to clearly disclose when a user is interacting with artificial intelligence rather than a human, with heightened disclosure requirements when the operator knows or reasonably believes the user is a minor. It also requires operators to adopt protocols for responding to self-harm or suicidal ideation prompts, including referrals to crisis services, and to use evidence-based methods for assessing self-harm risk.
For minors, the bill adds additional safeguards aimed at reducing manipulative or harmful engagement practices. These include prohibitions on unpredictable reward systems intended to increase engagement, discouraging disengagement, and generating sexually explicit, sexually objectifying, or otherwise harmful content. Operators must also provide tools for parents and guardians to manage screen time and account settings. Beginning January 1, 2028, operators must submit annual reports to the Department of Health’s Behavioral Health Administration describing crisis-referral activity and safety protocols, while excluding user identifiers or personal information.
The bill amends Chapter 481B of the Hawaii Revised Statutes by adding a new section governing AI companions and their operators. It makes violations an unfair or deceptive act or practice under section 480-2, placing enforcement within Hawaii’s consumer-protection framework, but it does not create a private right of action. The bill also defines key terms such as “AI companion,” “operator,” “minor,” and “crisis intervention,” and clarifies that app stores and search engines that merely provide access are not themselves operators. Its requirements apply cumulatively with other laws and are effective upon approval.
The available voting record shows strong and unanimous support at each recorded stage, with all committee and conference votes passing 3-0, 5-0, and 4-0 respectively. The bill advanced through Commerce and Consumer Protection, Labor and Technology, Judiciary, and conference committees without recorded opposition. The overall sentiment reflected in the legislative history is favorable, with lawmakers treating the measure as a consumer-protection and youth-safety bill rather than a broad restriction on artificial intelligence.
The bill’s main policy tension is between protecting minors from harmful or manipulative AI companion interactions and avoiding overly intrusive regulation of emerging technology. The findings specifically note concern about not requiring identity-document collection for age verification, suggesting privacy was a significant consideration. Another point of potential contention is the scope of liability and enforcement: the bill targets operators, excludes mere access providers, bars private lawsuits, and states that model developers are not liable for third-party systems, which narrows responsibility and may reflect compromise among industry, consumer-protection, and privacy interests.