OpenAI’s Sora 2 video generation app has taken the market by storm, reaching over 1 million downloads within just five days of its release. However, the company’s triumph has been clouded by serious controversy, as users rapidly produce deepfake videos featuring deceased celebrities and copyrighted characters. This surge has prompted fierce criticism from Hollywood studios and the families of those portrayed.
Launched as an invite-only iOS app available only in the U.S. and Canada, Sora 2 enables users to create hyper-realistic, short videos from text prompts. The app swiftly rose to the top of the Apple App Store charts, with its adoption rate reportedly outpacing the original launch of ChatGPT, according to OpenAI’s head of Sora, Bill Peebles.
In response, the Motion Picture Association (MPA) demanded that OpenAI take swift and strong measures to stop widespread copyright violations on the platform. Since Sora 2’s debut on September 30, users have generated unauthorized videos depicting characters from popular franchises. Viral examples include James Bond playing poker with OpenAI CEO Sam Altman and footage of Nintendo’s Mario fleeing police, all created without permission.
MPA CEO Charles Rivkin emphasized that copyright law places the responsibility on OpenAI to prevent infringement, dismissing OpenAI’s initial “opt-out” system where rights holders had to request exclusion of their intellectual property. Leading talent agencies WME, Creative Artists Agency, and United Talent Agency have collectively opted out of allowing their clients’ likenesses to be used on Sora 2.
Beyond intellectual property concerns, families of deceased celebrities have voiced distress over the app’s use of their loved ones’ images. Zelda Williams, daughter of the late Robin Williams, publicly pleaded for people to stop sending AI-generated videos of her father, condemning such creations as disrespectful distortions of human lives. Similarly, Kelly Carlin, daughter of George Carlin, described the experience as overwhelming, and the daughter of Malcolm X expressed pain at the casual use of her father’s likeness.
OpenAI has defended the depiction of historical figures as protected by strong free speech rights but allows authorized representatives of recently deceased individuals to request removal. However, the company has not clearly defined what qualifies as “recently deceased.”
Under increasing legal and public pressure, OpenAI CEO Sam Altman announced a policy shift from an opt-out to an opt-in system for copyrighted content generation. This change grants rights holders more precise control over character usage. However, users have found ways to bypass these protections through altered prompts, leaving some frustrated with the new restrictions.
Experts warn that the proliferation of hyper-realistic AI-generated videos using recognizable faces poses new challenges to online trust and digital safety. OpenAI is working on revenue-sharing models with rights holders and seeking more nuanced controls to balance innovation with legal and ethical considerations. Despite these efforts, industry insiders predict ongoing conflict as Silicon Valley’s fast-paced development clashes with Hollywood’s efforts to protect creative rights and enforce copyright law.
