Why Is AI Taking So Long to Make Its Mark in Hollywood?

In the early days of this year, companies like OpenAI and other AI startups made promises to Hollywood studios. They suggested that AI tools could speed up the filmmaking process, making it easier and cheaper to create movies and television shows.

What these tech companies sought was access to vast amounts of film footage and intellectual property. This data is essential for training AI models, which need quality content to generate new materials like video clips and script suggestions.

Despite the excitement, progress has been slow. One notable arrangement was recently revealed between New York-based AI startup Runway and Lionsgate, the studio behind popular franchises like “John Wick” and “Hunger Games.” As part of their partnership, Runway will develop a new AI model specifically for Lionsgate to enhance behind-the-scenes tasks, like storyboarding.

However, no other major studio has announced similar collaborations yet, and insiders speculate that these further partnerships may not materialize until 2025.

Several factors contribute to this delay. The AI landscape is complicated, marked by evolving regulations and legal uncertainties. Additionally, there is skepticism about whether audiences will embrace films predominantly created with AI tools. Questions about how to value studio libraries for AI training also loom large, alongside concerns regarding the protection of intellectual property.

In the entertainment industry, AI has sparked controversy largely due to mistrust of tech companies and their approaches toward intellectual property. Many creative professionals worry that AI applications, such as text-to-image and video tools, could threaten jobs within the industry.

The Lionsgate and Runway deal, for instance, triggered immediate worry among some creatives. Following the announcement, many professionals reached out to their agents for clarification. The Directors Guild of America has already contacted Lionsgate and plans to discuss the implications of this agreement soon.

Fears surrounding AI were a crucial issue raised by actors and writers during the strikes last year. They pushed studios to include protections in contracts, such as requiring permission from actors to create digital likenesses and establishing compensation models for their use. Any future agreements will likely have to consider these restrictions.

“Companies are proceeding cautiously from a legal standpoint, trying to figure out how to handle AI in a responsible way,” said Javi Borges, head of a team advising entertainment and media firms for professional services company EY. “While AI is currently a buzzword, many are still navigating how to maximize its potential benefits.”

The Alliance of Motion Picture and Television Producers, representing major studios in labor negotiations, chose not to comment on this topic. Studios like Netflix, Walt Disney Co., and Warner Bros. Discovery either declined to comment or didn’t respond to inquiries.

Amid ongoing streaming challenges and a recovering theatrical market post-COVID-19, studios find AI appealing as a potential cost-cutting measure.

Recently, Warner Bros. Discovery announced plans to utilize Google’s AI technology for captioning unscripted programming, stating this would significantly reduce production time and costs. Reports also indicate that Warner Bros. Discovery and Disney are exploring partnerships with OpenAI to license footage from their libraries.

Additionally, Meta has been working with Blumhouse, a well-known horror producer based in Los Angeles, on a pilot project meant to gather feedback from creative professionals on their Movie Gen AI tools. Blumhouse involved three filmmakers, including Casey Affleck, to experiment with AI-generated clips, allowing them to express creative ideas more rapidly.

One major obstacle to establishing more partnerships is the lack of a standardized valuation for what film and TV libraries are worth to AI companies. Hollywood typically values a studio’s content based on popularity, while AI firms are interested in a wide range of data for model training, meaning that even lesser-known works could hold significant value.

“Forward-thinking executives at companies with valuable intellectual property understand the importance of their assets in the long term,” said Dan Neely, co-founder and CEO of Vermillio, an AI rights management platform. “However, without a pricing standard, this is a pivotal moment for major IP holders to get right.”

Legal issues surrounding how AI models learn and how creators should be compensated remain significant challenges. Lawsuits alleging that works were improperly used to train AI models are already emerging against companies like Runway. Many music labels and publishers are also pursuing legal action against businesses in the AI sector.

There’s concern within Hollywood regarding OpenAI’s view on copyright. The company holds an expansive interpretation of fair use, allowing for limited use of copyrighted material without the owner's permission. While OpenAI offers an opt-out for entities wishing to avoid AI training, that does little to alleviate fears for copyright holders.

In May, the situation escalated when Scarlett Johansson accused OpenAI of using a chatbot that mimicked her voice, which she had previously rejected. OpenAI defended its position but ultimately took the chatbot down.

“Potential partnerships are overshadowed by copyright uncertainties and producers’ desire to maintain the tradition of human-centered storytelling,” stated Duncan Crabtree-Ireland, national executive director and chief negotiator for SAG-AFTRA, via email.

While major studios have not yet initiated legal actions against AI companies, some industry analysts hint that such moves may lie ahead.

On another front, state and federal lawmakers are working to implement legislation to address public concerns about AI. Recently, California’s Governor Gavin Newsom signed laws aimed at tackling the spread of deep fakes in political advertising, but many in Hollywood believe more actions are necessary.

The Lionsgate and Runway partnership offers insight into how future collaborations with studios might be structured. In their agreement, Runway will develop a model trained on a select few titles, aiding Lionsgate in improving production and marketing efficiency. Importantly, this model will not aim to replicate actors’ likenesses, and the specific data will not be used to train Runway’s other models.

No financial transactions were reported in the partnership, as both parties opted not to comment on the details of the deal.

Runway’s CEO, Cristóbal Valenzuela, remarked, “We are all realizing how powerful this technology can be, but it isn't a magic solution that creates films independently. Instead, these tools serve as valuable assets for artists. Many industry professionals are beginning to embrace and understand the potential changes for their craft.”

As companies continue discussions, AI startups are actively seeking to connect their tools with creators, sometimes offering financial incentives. Several creators are already integrating AI into their work for music videos and enhancing content for platforms like YouTube.

However, the speed at which this will lead to wider acceptance and agreements within major studios remains uncertain. An anonymous executive shared, “Ultimately, it will come down to financial incentives. If the offer is substantial enough, studios are likely to overlook initial concerns and navigate the complexities surrounding profit sharing.”

AI, Hollywood, Partnership, Lionsgate, Runway