OpenAI has released a joint statement with SAG-AFTRA (the Screen Actors Guild – American Federation of Television and Radio Artists), actor Bryan Cranston, and major Hollywood agencies, confirming plans to strengthen content guardrails in its Sora 2 video platform to address unauthorized likeness and voice replication. The collaboration follows reports of AI-generated videos featuring celebrities without consent, reigniting debates about consent and digital personhood.
The controversy began when Cranston discovered AI-generated clips of himself circulating on Sora 2 — including one depicting him taking selfies with Michael Jackson. OpenAI issued an apology for these “unintentional generations” and pledged to work with the entertainment industry on a framework for permission-based participation. The company described the new measures as part of a “productive collaboration” with SAG-AFTRA and Hollywood stakeholders.
Why it matters: The move marks a significant turning point for AI entertainment ethics. Sora 2 has seen explosive adoption among creators — but also backlash from actors, estates, and regulators over unauthorized use of likenesses. The new safeguards aim to balance creative freedom with consent, signaling a maturing phase for generative video. With prominent voices like Cranston, Sean Astin, and Martin Luther King Jr.’s estate speaking out, Hollywood’s relationship with AI may now hinge on how effectively platforms like Sora enforce these promises.
Read the full statement from SAG-AFTRA: SAG-AFTRA and OpenAI: Protecting Voice and Likeness on Sora 2.