AI-Generated Videos of Stars Spark Copyright Clash in China’s Tech Race
A 15-second AI-generated video depicting a fight between Tom Cruise and Brad Pitt has triggered a wave of concern in Hollywood, prompting legal action and sparking debate about the future of filmmaking. The video, created using a latest AI tool called Seedance 2.0 developed by Chinese tech company ByteDance, quickly went viral, demonstrating a level of realism that alarmed industry professionals.
The clip, showing the two actors engaged in a rooftop brawl, was created “simply by entering a two-line prompt,” according to Irish filmmaker Ruairi Robinson, who posted it on X (formerly Twitter). The ease with which the video was produced and its convincing quality, led Deadpool & Wolverine screenwriter Rhett Reese to declare on X, “I hate to say it. It’s likely over for us.” He later elaborated, stating that someone with the talent of Christopher Nolan could soon create entire movies indistinguishable from those produced by major studios.
The rapid advancement of AI video generation has prompted immediate responses from Hollywood’s legal and trade organizations. Paramount and Disney sent cease and desist letters to ByteDance, accusing the company of copyright infringement, according to reports. The Motion Picture Association (MPA) and the actors’ union SAG-AFTRA also condemned ByteDance for unauthorized apply of copyrighted American works. ByteDance responded with a statement promising to implement stronger safeguards to protect intellectual property.
Seedance 2.0 is the latest in a series of AI models released by Chinese tech companies, as competition intensifies to dominate the artificial intelligence industry. The Chinese government has prioritized advanced technology as a key component of its national development strategy, showcasing robotics and AI capabilities during recent public events, including the Lunar New Year celebration.
The emergence of this technology has fueled anxieties in the United States, mirroring the “space race” of the 20th century with the Soviet Union. Ramesh Srinivasan, a professor of information studies at UCLA, noted a “fervor” surrounding which nation will lead in AI development.
The tool’s capabilities extend beyond simply creating realistic visuals. A Chinese tech blogger reportedly demonstrated that Seedance 2.0 could generate audio mimicking their voice based solely on an image, raising concerns about deepfakes and privacy. ByteDance subsequently removed this feature and introduced verification requirements for users creating digital avatars, according to Chinese media reports.
Experts suggest that China’s rapid release of new AI iterations may be outpacing its ability to assess and mitigate potential negative consequences. Rogier Creemers, a lecturer at Leiden University specializing in Chinese technology policy, cautioned that more powerful applications inherently carry greater risks, comparing the situation to developing a faster car – increased speed comes with a higher risk of accidents.
Following the backlash, ByteDance stated its commitment to respecting intellectual property rights and strengthening protections against unauthorized use of copyrighted material. The popular Chinese social media app RedNote also responded by requiring AI-generated content to be clearly labeled. These actions coincide with a tightening of regulations surrounding AI-generated content within China itself.
China’s regulations on AI content are among the most stringent globally, owing to its established censorship apparatus. The Cyberspace Administration of China recently announced a crackdown on unlabeled AI-generated content, penalizing over 13,000 accounts and removing hundreds of thousands of posts. Still, enforcement of these regulations has been inconsistent, according to Nick Corvino of ChinaTalk, a newsletter focused on Chinese affairs, due to competition among platforms and incentives to maximize user engagement.
Analysts suggest China is navigating a delicate balance between fostering domestic AI innovation and maintaining control over its application. Whereas strict content controls may hinder development, they could also spur alternative approaches, as demonstrated by the Chinese company DeepSeek, which built a competitive chatbot with a smaller budget.
The situation also highlights the importance of data access for AI model training. Disney recently reached an agreement with OpenAI, granting Sora – a competitor to Seedance 2.0 – access to characters like Mickey and Minnie Mouse, suggesting that access to high-quality data is a key factor in AI advancement.