Scripts from more than 130,000 films and TV shows have been used to train generative AI models, which a BFI report said today “poses a threat to the fundamental economics of the screen sector.”

The report taps into major concerns over artificial intelligence and copyright, coming as the UK government considers legislation that would mean copyright holders must opt-out from having their material used for training AI models, which has caused consternation amongst producers and rightsholders.

The scale of the problem was communicated by the BFI’s AI in the Screen Sector: Perspectives and Paths Forward, which said more than 130,000 film and TV scripts have been used to train AI models.

The report called copyright “the dominant concern around AI today” and said many models “in wide use have been trained on copyrighted material without the permission of rightsholders or any form of payment to creators.”

“As generative models learn the structure and language of screen storytelling – from text, images and video – they can then replicate those structures and create new outputs at a fraction of the cost and expense of the original works,” added the report. “These learned capabilities can be used to assist human creatives, but AI tools may also be used to compete against the original creators whose work they were trained on.”

The report added that the “existing training paradigm for generative AI models poses a threat to the fundamental economics of the screen sector and its ability to create value from making and commercialising new IP.”

The opt-out legislation is currently making its way through parliament but has been heavily criticized. Speaking last week at a Deloitte media conference, Sky CEO Dana Strong said “some consequences of the opt out are impossible to police.”

“Duality” of excitement and scepticism

Away from copyright concerns, today’s report floated a “duality” of both “excitement and scepticism” that “characterises much of the industry discourse surrounding generative AI.”

While the review pointed to high-profile examples of AI-induced backlash such as Oscar-winner The Brutalist’s enhancement of the pronunciation of Hungarian dialog, it added that “screen sector adoption of generative AI is also happening behind the scenes, without public outcry.”

A 2023 survey reported that 17% of UK producers had used AI, with 40% planning to do so, while 2024 surveys found that nearly half of U.S. media executives and 40% of French screen professionals were using AI. When it comes to the BFI, the institute reported that approximately 8% of more than 3,000 UK-based fund applicants reported using AI to complete their submissions over the past 18 months.

In general, generative AI is “not perfect but is improving in its suitability for creative tasks,” the report said.

“The stakes are high,” wrote BFI research director Rishi Coupland in the report’s foreword. “Without strategic planning, the UK screen sector may find itself outpaced by global competitors and new AI-native studios. The sector’s future may depend on its ability to harness AI’s benefits while mitigating its risks.”