Concerns have emerged among actors, studios, agents, and the SAG-AFTRA union regarding the AI-generated content produced by Sora 2, particularly since the platform’s launch last month. In response to these apprehensions, a joint statement was issued by actor Bryan Cranston, OpenAI, and the union, indicating that the company has implemented “strengthened guardrails” around its opt-in policy following instances where videos featuring Cranston, including one where he appeared to take a selfie with Michael Jackson, were generated without consent.
The statement also noted that OpenAI expressed regret for these inadvertent outputs and received support from major talent agencies, such as the United Talent Agency and the Association of Talent Agents, which have previously criticized the lack of protections for artists. However, OpenAI has not provided detailed information regarding specific changes to the Sora platform or addressed requests for comment from media outlets.
Furthermore, OpenAI affirmed its commitment to ensuring that all artists and individuals retain the right to control how their likenesses and voices are utilized in simulations. The company stated it would prioritize the review of complaints regarding policy violations.
While Cranston acknowledged his appreciation for OpenAI’s policy improvements, SAG-AFTRA president Sean Astin highlighted the need for legislative measures to protect performers from unauthorized use of their likenesses, referencing a proposed bill known as the NO FAKES Act, aimed at addressing these issues.
Initially, OpenAI’s Sora 2 launched with an opt-out policy for rights holders, but due to public backlash over problematic content, including controversial AI-generated figures, it shifted to a model promising enhanced control for creators, akin to the opt-in system for likeness usage.
Source: https://www.theverge.com/news/803141/openai-sora-bryan-cranston-deepfakes