The measure would bar some knowing, unauthorized uses of lifelike AI imitations and let depicted people seek court orders or damages, with exceptions for protected expression.
The measure would bar knowingly publishing a lifelike AI imitation without written consent when it appears in an ad, causes harm, or is used for certain crimes. Several forms of expression, including satire and news illustrations, are exempt.
A person depicted in a qualifying imitation could seek its removal and money damages. An estate could sue for up to ten years after the person's death; punitive damages require proof of malice.
The measure would make it unlawful to knowingly publish a lifelike AI-made imitation of an identifiable person without written consent in certain cases. It covers a lifelike AI-made depiction, audio, or video based on an actual, particular person who did not perform or appear in it. The imitation must be used in an ad, cause harm, or be used for fraud, defamation, harassment, or another crime. Exemptions cover parody, satire, commentary, criticism, scholarship, political or educational expression, news illustrations, and documentary, historical, or biographical works. Ads for those works are exempt only when the imitation relates to the work. The act says it takes effect upon approval; the supplied status conflicts with the signed act about whether approval occurred.
A depicted person or their estate could seek a court order to remove or stop sharing a qualifying imitation, along with money damages and other remedies. An estate could bring an action for up to ten years after the depicted person's death. A lawsuit is available only for a nonexempt publication that violates the consent rule. Damages may be up to $25,000 per ad or actual damages. Punitive damages require proof of malice, and the usual filing time limits still apply. The ten-year period applies after the depicted person's death. Other lawsuit filing time limits still apply.
No hearings or votes scheduled.