Hawaii Signs New AI Laws Covering Deepfakes and AI Companions
Hawaii’s new AI laws took effect July 14, creating conditional civil remedies for qualifying deepfakes and disclosure, crisis-response and minor-protection rules for AI companion operators.
Hawaii has enacted two statewide laws regulating high-risk uses of artificial intelligence, giving residents conditional civil remedies for certain unauthorized digital imitations and imposing disclosure, crisis-response and minor-safety duties on covered AI companion operators.
Gov. Josh Green signed House Bill 2137 as Act 247 and Senate Bill 3001 as Act 248 on July 14, 2026. Both laws took effect upon approval, according to the final enacted texts. The laws do not ban all AI-generated images, videos or conversations. Their requirements apply only when the statutory conditions are met.
Act 247 targets harmful unauthorized digital imitations
Act 247 covers a “realistic digital imitation” of an identifiable individual. The law defines that as a highly realistic AI-generated depiction, audio or video based on an actual person’s identifying characteristics and portraying that person’s voice, face or likeness in a sound recording or audiovisual work in which the person did not actually perform or appear.
It is unlawful for a person to knowingly publish such an imitation without the individual’s consent when the imitation is used in connection with an advertisement, causes harm, or is used to commit fraud, defamation, harassment or another criminal act.
The law defines harm to include reputational injury, financial loss, emotional distress or the misappropriation of identity for commercial gain. Consent must be express, written permission from the depicted individual or the individual’s authorized representative.
What remedies may be available
A person whose qualifying digital imitation was published may bring a civil action if the content violated Act 247 and was not covered by one of the law’s exemptions. The person’s estate may also bring an action for up to 10 years after the individual’s death, although the law does not extend any otherwise applicable statute of limitations under Hawaii’s existing law.
Available remedies may include an injunction ordering the publisher to remove the imitation or stop distributing it. A plaintiff may seek monetary damages of up to $25,000 per advertisement or may seek actual damages, including damages for reputational injury and emotional distress. Punitive damages are available where malice is proven. The law also allows reasonable attorneys’ fees and court costs.
Those remedies are not automatic. A claimant still must meet the law’s conditions and prove the facts required in a civil proceeding. The final text also preserves other remedies that may be available under separate laws.
Speech and third-party publication exemptions
Act 247 exempts AI-generated or AI-altered content that is parody, satire, commentary, criticism, scholarship, or political or educational expression. It also exempts news reporting when the content is used to illustrate a story, as well as certain documentary, historical and biographical portrayals. Advertising for those exempt works is covered when the digital imitation is relevant to the work being promoted.
The law separately protects a medium that publishes or disseminates third-party content, including newspapers, television stations, streaming services, cable systems and transit advertisements. That protection applies only to the medium’s role in distributing the third-party material and does not exempt the underlying conduct regulated by the act.
The Hawaii attorney general may bring an action for injunctive or other equitable relief when distribution of realistic digital imitations involves broad public interest or widespread harm. That authority does not eliminate an individual’s or estate’s right to pursue a civil action.
Act 248 regulates AI companion services
Act 248 applies to operators that develop and make AI companions available to the public. The law defines an AI companion as a system designed to simulate a sustained human or human-like relationship by retaining prior interactions or preferences, asking unprompted emotion-based questions and maintaining an ongoing dialogue about personal matters.
If a reasonable person could believe they are interacting with a human, the operator must provide a clear and conspicuous notice that the companion is artificial intelligence and not human. When an operator has actual knowledge or reasonable certainty that a user is a minor, the disclosure must be persistent or appear at the beginning of each session and at least once every hour during a continuous interaction. The hourly notice must also remind the user to take a break.
Operators must adopt protocols for prompts involving suicidal ideation or self-harm. Those protocols must include reasonable efforts to refer users to crisis-service providers, use evidence-based methods to measure suicidal ideation and self-harm risk, and prevent the AI companion from representing that it is designed to provide professional mental or behavioral health care.
Operators also must take reasonable measures to prevent an AI companion from making statements that could lead a reasonable person seeking or receiving crisis intervention to believe they are speaking with a human, or from generating outputs that encourage a user to cause serious bodily injury to another person.
Additional protections for minors
When an operator knows or has reasonable certainty that a user is a minor, Act 248 bars the operator from providing points or similar rewards at unpredictable intervals when intended to increase engagement with the AI companion. It also prohibits the AI companion from generating outputs that discourage the minor from disengaging.
Operators must take reasonable measures to prevent the system from producing visual material of sexually explicit conduct, directly telling a user to engage in sexually explicit conduct, or making statements that sexually objectify the user. They must also make tools available for users and their parents or guardians to manage screen time and account settings.
Enforcement and what happens next
Violations of Act 248 are treated as unfair or deceptive acts or practices under Hawaii’s consumer-protection law. The final text does not create a private right of action to enforce that section or to support a private right of action under another law. It also states that the law does not create liability for the developer of an AI model when a third party uses that model to provide an AI companion.
Beginning January 1, 2028, operators must submit annual reports to the Hawaii Department of Health’s Behavioral Health Administration. The reports must identify the number of crisis-service referrals issued during the preceding calendar year and describe protocols for detecting, removing and responding to prompts involving suicidal ideation or self-harm, including protocols intended to prohibit responses promoting suicidal ideation, suicide or self-harm. The reports may not include user identifiers or personal information.
For residents, the immediate change is that Hawaii law now provides a framework for challenging certain unauthorized synthetic media and sets statewide operating requirements for covered AI companion services. The practical availability of relief will depend on the facts of each case, the statutory exemptions, enforcement decisions and any court proceedings that follow.
Sources
- HB 2137, C.D. 1 final bill text
- Gov. Green Signs Legislation to Support Kūpuna Care and Strengthen AI Protections
- Experts praise Hawaii’s new laws on abusive use of generative AI
Discover more from Interactive News
Subscribe to get the latest posts sent to your email.
