MPA Demands ByteDance Restrain AI Model Generating Deepfakes of Tom Cruise and Brad Pitt

The Motion Picture Association (MPA) has issued a formal demand to ByteDance, the Chinese tech giant behind TikTok, to immediately curb the use and dissemination of a newly developed artificial intelligence model capable of generating hyper-realistic deepfakes of A-list celebrities—including Tom Cruise and Brad Pitt—from copyrighted motion picture content. According to Deadline, the model, internally referred to as 'Seedance,' allows users to extract and manipulate facial features and vocal patterns from existing films and TV shows, producing convincing synthetic performances that mimic the original actors without consent or licensing.

While the MPA has not publicly disclosed the model’s origin, it is believed to have been developed by an independent research group or rogue developer and subsequently integrated into third-party applications accessible via TikTok’s ecosystem. The MPA’s statement emphasizes that the AI’s ability to replicate copyrighted performances constitutes a direct violation of intellectual property rights under U.S. and international copyright law. The association warned that such tools, if left unchecked, could undermine the economic foundations of the film industry by enabling unauthorized commercial use of actors’ likenesses and eroding the value of original content.

Deepfakes featuring Tom Cruise have previously surfaced on social media platforms, notably in a viral 2024 video that convincingly depicted him delivering a motivational speech in a style mimicking his public persona. That video, later attributed to a separate AI tool, was removed following public outcry and legal pressure. However, the 'Seedance' model represents a significant escalation: it is reportedly trained on proprietary studio archives, allowing it to generate deepfakes from a wider range of films—including those not publicly available—raising alarms among major studios about the potential for insider data leaks or unauthorized training data acquisition.

ByteDance has not yet issued a public response to the MPA’s demand, but internal sources suggest the company is evaluating the model’s presence on its platforms. TikTok’s community guidelines prohibit the use of AI to deceive or mislead users, particularly regarding identity and authenticity. However, enforcement remains inconsistent, especially when content is distributed through encrypted channels or masked as "educational" or "artistic" experiments. The MPA has urged ByteDance to implement real-time AI detection tools and to collaborate with studios on a shared database of protected likenesses to prevent future misuse.

Legal experts warn that this case could set a precedent for how AI-generated content is regulated under existing copyright frameworks. While current laws protect the reproduction of copyrighted works, the legal status of synthetic performances using publicly available footage remains ambiguous. The MPA is now lobbying Congress and international bodies to clarify whether AI models trained on copyrighted material without permission constitute infringement, even if the output is not a direct copy.

Meanwhile, actors’ unions, including SAG-AFTRA, have voiced strong support for the MPA’s stance. "This isn’t just about copyright—it’s about consent," said a union spokesperson. "No actor should have their face, voice, or performance weaponized by an algorithm they never agreed to." The controversy has reignited debates over AI ethics in entertainment, with calls for mandatory disclosure of synthetic media and new legislation to protect digital identity.

As the industry braces for a wave of similar AI tools, the MPA’s intervention signals a turning point in the battle between technological innovation and creative rights. Whether ByteDance will comply—or whether the model will proliferate across decentralized networks—may determine the future of digital authenticity in Hollywood and beyond.