Lionsgate's AI Film Push Hits Snags as Copyright and Tech Limits Emerge

This article was compiled by the editorial desk based on reporting from The Wrap, Vulture, and public statements from Lionsgate executives and legal experts.

LOS ANGELES — A year after Lionsgate announced a landmark partnership with AI startup Runway to develop a custom video generation model, the initiative is struggling with technical limitations and unresolved copyright questions, according to industry reports. The studio, which granted Runway access to its extensive film library—including franchises like The Hunger Games and American Psycho—hoped to create a tool capable of producing "cinematic video" for its exclusive use. However, insiders say the model's output has fallen short of expectations, and legal uncertainties loom over the entire endeavor.

In an interview with Vulture last month, Lionsgate vice chairman Michael Burns touted the potential of the AI model, suggesting it could generate reimagined versions of the studio's properties, such as an anime-style John Wick. But according to a report from The Wrap, the project has encountered significant obstacles, including a shortage of quality training data and the inherent limitations of current AI technology. A source familiar with the matter told The Wrap, "The Lionsgate catalog is too small to create a model. In fact, the Disney catalog is too small to create a model."

The issue extends beyond Lionsgate. Disney's attempt to use AI to deepfake Dwayne "The Rock" Johnson's face into the upcoming live-action Moana sequel also failed. After 18 months of work with the firm Metaphysic, none of the AI-assisted scenes will appear in the final film, according to the report.

These setbacks highlight a broader industry challenge: the scarcity of high-quality training data. Even scraping the entire internet has not produced a refined AI video tool, and fine-tuning on a studio's catalog has not been enough to overcome the technology's current limitations. The result is output that, while impressive in short clips, lacks the coherence and visual language of traditional cinema, often resembling a montage of b-roll rather than a cohesive narrative.

Copyright issues further complicate the picture. The generative AI industry operates in a legally gray area, with many models trained on data without explicit permission. Recent legal actions, such as Anthropic's $1.5 billion settlement with authors for using their books in training, underscore the risks. Jacob Noti-Victor, a professor at Cardozo Law School, told The Wrap that even if studios own the intellectual property that AI adapts from, the AI-generated work itself may not receive full copyright protection. "I think the studios would be leaning on the fact that they would own the IP that the AI is adapting from, but the work itself wouldn't have full copyright protection," he said.

Why the AI Film Dream Remains Distant

The fundamental problem, experts say, is that AI video generators are not yet capable of producing entire feature films. While they can generate short sequences, the technology struggles with consistency and narrative flow. The uncanny, aesthetically amorphous output is a far cry from the polished, intentional visuals audiences expect from Hollywood productions. AI is already being used to enhance or replace visual effects in some cases, but generating a full movie remains out of reach.

For Lionsgate, the promise of an AI-generated anime John Wick will have to wait. The studio's bold bet on AI, while forward-thinking, has collided with the reality of the technology's current state. As the industry grapples with these challenges, the future of AI in filmmaking remains uncertain, but the potential—and the pitfalls—are becoming clearer.