OpenAI is tightening safeguards around Sora 2 after actor Bryan Cranston's voice and likeness appeared in generated videos without his consent. The incident violated the company's official opt-in policy and has become another flashpoint in the debate over AI video, performers' rights, and digital identity.
What Changed After The Cranston Incident
OpenAI described the videos as "unintentional generations" and said it has "strengthened guardrails around replication of voice and likeness" for Sora 2. The change followed talks with SAG-AFTRA and Cranston, who reported the incident to the union.
The company says the opt-in policy remains in place. Under that policy, artists are supposed to retain full control over how their digital likeness is used.
Cranston framed the issue as bigger than one actor. "I was deeply concerned not just for myself, but for all performers whose work and identity can be misused in this way," he said.
OpenAI has also promised to handle complaints quickly. Cranston welcomed the changes, while stressing that control over one's voice and appearance is a fundamental right for all artists.
Why The Opt-In Policy Matters
The central issue is consent. If an AI video generator can produce a recognizable version of a person without permission, a policy that promises artist control becomes meaningful only if the system can reliably prevent unauthorized output.
In this case, the source says Cranston's voice and likeness appeared even though the policy required opt-in permission. That gap is why the company's response focused on stronger guardrails, not on replacing the policy itself.
For performers, the risk is direct. A voice, face, and recognizable appearance are part of the work. If those traits can be replicated without approval, the concern is not only reputational, but also professional and personal.
The same logic applies beyond a single performer. OpenAI's response, SAG-AFTRA's involvement, and Cranston's comments all point to a broader conflict over whether AI systems can be deployed quickly while still protecting the people whose identities they may imitate.
A Joint Declaration Puts More Pressure On AI Video
The new safeguards are part of a joint declaration signed by SAG-AFTRA, OpenAI, Cranston, United Talent Agency, Creative Artists Agency, and the Association of Talent Agents (ATA). That lineup shows how the issue reaches across performers, unions, agencies, and the company behind the tool.
The group also backs the NO FAKES Act, proposed federal legislation that would ban unauthorized digital copies of a person's voice or likeness. The source does not provide further details about the proposal, but its inclusion shows that the dispute is moving beyond product policy and into the legal arena.
Cranston has been an outspoken critic of AI's role in Hollywood since 2023. His position has focused on respect for actors' rights and warnings against replacing human labor with AI.
The declaration therefore serves two purposes. It announces a practical response to the Sora 2 incident, and it links that response to a wider push for stronger protection against unauthorized digital copies.
Sora 2 Faces A Broader Safeguards Test
The Cranston incident did not happen in isolation. Since its launch in early October, Sora 2 has faced steady criticism for generating unauthorized imitations of celebrities and copyrighted content, including entire episodes of "South Park."
The source argues that OpenAI leaned into launch hype at the expense of strong safeguards, only introducing restrictions after public backlash. It also compares the approach to OpenAI's new image model in ChatGPT, citing the Ghibli case.
CEO Sam Altman has promised that rights holders will get a share of revenue, though the details are still unclear. That uncertainty matters because revenue sharing and consent are related but not identical questions. Payment does not automatically resolve whether someone approved the use of a voice, face, style, or likeness in the first place.
A recent NewsGuard investigation found that Sora 2 can generate convincing fake videos in minutes with minimal effort. The source describes this as a striking change for OpenAI, which once withheld its relatively weak language model GPT-2 out of concern for fake news, but now ships a video model that NewsGuard warns could be used to spread disinformation.
The Stakes For Artists And Platforms
The immediate outcome is clear: OpenAI says it has strengthened Sora 2 safeguards, kept its opt-in policy, and committed to faster complaint handling. Cranston has welcomed those changes while keeping the focus on artist control.
The larger question is whether safeguards can keep pace with a system that can generate realistic video and recognizable identities. The source presents Sora 2 as a product under pressure from multiple directions: celebrity likeness concerns, copyrighted content, disinformation risk, and unclear revenue-sharing details.
For OpenAI, the challenge is no longer simply whether Sora 2 can generate compelling video. It is whether the company can prove that consent, complaints, and rights-holder protections are built into the product strongly enough to prevent the next misuse before it happens.