Lawyers say OpenAI could be in real trouble with Scarlett Johansson

OpenAI has potentially landed itself in hot water by creating a voice for its ChatGPT that mimics the voice of actress Scarlett Johansson, a similarity noted by both the public and media during a product demo. Intellectual property lawyers have expressed that Johansson has a strong case under California's right to publicity laws that prevent the unauthorized commercial use of an individual's likeness, including their voice. These laws are particularly rigorous in cases where a celebrity's distinctive voice is used in a way that might imply their endorsement or involvement without their consent.

The issue was compounded when Sam Altman, CEO of OpenAI, made a cryptic post online that seemed to acknowledge the similarity during the demo, possibly influencing public perception that the AI voice was intentionally modeled after Johansson. Such actions by OpenAI come after Johansson had already turned down a request from the company to officially voice the assistant, indicating that her distinctive voice was known and acknowledged by OpenAI before the creation of the Sky voice. Legal precedents like those involving celebrities Bette Midler and Tom Waits, who successfully sued companies for unauthorized use of vocal likenesses, suggest Johansson could prevail should she decide to litigate.

While OpenAI has temporarily pulled the controversial voice from its platform, the matter underscores ongoing legal and ethical challenges faced by companies working with AI in creating lifelike digital personas. The absence of cohesive federal right to publicity laws only adds complexity, leaving largely uncharted territory about digital replicas and AI. If Johansson proceeds with legal action or a settlement is not reached, this case could set important precedents for how AI companies use celebrity likenesses in the future.