OpenAI cracks down on Sora 2 deepfakes after pressure from Bryan Cranston, SAG-AFTRA

In a significant move to address concerns over deepfakes, OpenAI has announced its collaboration with actor Bryan Cranston, SAG-AFTRA, and various other unions to enhance protections on its AI video creation platform, Sora. This initiative follows the emergence of unauthorized AI-generated clips that utilized Cranston's voice and likeness shortly after the launch of Sora 2 at the end of September. Cranston, known for his roles in "Breaking Bad" and "Malcolm in the Middle," voiced his worries about these unauthorized uses, prompting a statement from SAG-AFTRA on social media platform X. "I appreciate OpenAI's proactive approach in refining its policies and guardrails, and I hope all companies involved will respect our rights regarding the use of our voices and likenesses," he noted. To fortify the defenses against unapproved AI creations, OpenAI has pledged to work alongside United Talent Agency, which represents Cranston, as well as the Association of Talent Agents and Creative Artists Agency. These organizations have previously criticized OpenAI for allegedly misusing copyrighted materials, labeling Sora as a potential threat to their clients' intellectual property. Recently, OpenAI took action to block videos featuring Martin Luther King Jr. on Sora, following requests from King's estate due to the appearance of disrespectful portrayals. Additionally, Zelda Williams, the daughter of the late comedian Robin Williams, appealed for an end to AI-generated videos of her father soon after Sora 2's release. Since the launch of Sora 2 on September 30, OpenAI has adjusted its copyright policies. On October 3, CEO Sam Altman announced modifications to Sora's opt-out policy, which was previously more permissive regarding the use of intellectual property. Now, rightsholders will have increased control over the representation of their characters. OpenAI also committed to responding promptly to any complaints. The company reiterated its backing of the NO FAKES Act, a federal initiative aimed at safeguarding against unauthorized AI-generated reproductions of individuals' voices or likenesses. "OpenAI is committed to preserving the rights of performers against the misappropriation of their identities," Altman stated. "We were early advocates of the NO FAKES Act when it was presented last year and remain steadfast in our support for the rights of artists."