RELATING TO ARTIFICIAL INTELLIGENCE.
HB2137 creates a new Hawaii law aimed at restricting harmful uses of artificial intelligence-generated “digital imitations,” commonly known as deepfakes, when they realistically portray a real person’s voice, face, or likeness without consent. The bill makes it unlawful to knowingly publish such an imitation if it is used in an advertisement, causes harm, or is used to commit fraud, defamation, harassment, or other criminal acts.
The measure defines key terms such as artificial intelligence, consent, realistic, publish, and digital imitation, and it sets out exemptions for parody, satire, commentary, criticism, scholarship, political or educational expression, news reporting, and documentary, historical, or biographical uses. It also limits application to the fullest extent allowed by federal law and preserves existing federal protections for online platforms under Section 230.
The bill adds a new chapter to the Hawaii Revised Statutes regulating AI-generated realistic depictions of identifiable individuals. It creates a private right of action for the affected person or the person’s estate, allowing injunctive relief, removal or cessation of distribution, actual damages or up to $25,000 per advertisement, punitive damages if malice is shown, and attorneys’ fees and costs. It also authorizes the attorney general to seek equitable relief in cases involving broad public interest or widespread harm, while carving out exemptions for certain media intermediaries and preserving existing federal law limits.
The bill appears to have broad support and little recorded opposition. It passed multiple Senate committees and conference committees unanimously or near-unanimously, with votes such as 3-0, 4-0, and 5-0, and was ultimately transmitted to the Governor. The legislative findings frame the measure as a response to growing concerns about deepfake misuse, identity theft, fraud, and emotional harm, while also emphasizing protection for lawful speech and journalism.
The main policy tension in the bill is between preventing harmful deepfakes and avoiding overreach that could burden protected expression. The legislature specifically addressed this by exempting parody, satire, criticism, scholarship, political and educational expression, news reporting, and documentary or biographical uses. Another point of concern is platform and media liability, which the bill narrows by excluding certain third-party content distributors from direct application and by stating that it does not alter Section 230 protections. The civil damages provision and the scope of what counts as a “realistic” imitation could also be areas of concern, but the available vote history shows no recorded committee dissent.