Back to Timeline

Event Summary

On May 20, 2024, actress Scarlett Johansson publicly challenged OpenAI over a ChatGPT voice called 'Sky' that sounded 'eerily similar' to her own. Johansson revealed that OpenAI had approached her nine months earlier to license her voice, she declined, and the company used a voice she believed imitated hers anyway. OpenAI denied deliberate imitation but suspended the voice. The incident became a landmark moment in the debate over AI consent, celebrity likeness rights, and the ethical boundaries of AI voice synthesis.

Context & Narrative

The controversy erupted immediately after OpenAI's GPT-4o demo on May 13, where the 'Sky' voice demonstrated flirtatious, emotionally expressive conversation — drawing instant comparisons to the 2013 film 'Her', in which Scarlett Johansson voiced an AI assistant. When journalists and the public pointed out the similarity, Johansson released a statement saying she was 'shocked, angered and in disbelief' that OpenAI used a voice so similar to hers after she had declined their offer. The timing was particularly charged: the film Her is about a man falling in love with an AI assistant voiced by Johansson, and OpenAI CEO Sam Altman had tweeted a single word — 'her' — on the day of the GPT-4o demo. OpenAI denied the imitation, stating that Sky's voice belonged to a different professional actress using her natural speaking voice. However, the company suspended Sky's use. The incident had lasting effects: it intensified the debate around AI voice consent, influenced California's consideration of AI likeness protection laws (later part of SB 53), and became a cautionary tale cited in virtually every subsequent AI voice ethics discussion. It also highlighted the power imbalance between individual creators and AI companies.

Key Findings

  • Fact Grade A

    Scarlett Johansson publicly challenged OpenAI on May 20, 2024, stating that ChatGPT's 'Sky' voice sounded 'eerily similar' to hers after she had declined OpenAI's licensing offer. OpenAI suspended the voice.

    Sources [1][2]

Impact Assessment

  • Risk Creation -2 · Medium-term

    Demonstrated the power imbalance between AI companies and individuals whose voices/likenesses could be replicated without consent. Raised fundamental questions about AI voice consent, celebrity likeness rights, and the adequacy of existing legal protections.

    Affected Groups: celebrities, voice actors, general public, AI companies

Consensus & Sources

Significance L1
Category Safety & Ethics / Society & Culture
Consensus Broad Consensus
Impact Index 5/10