Edge

EXCLUSIVE: A Television Academy task force developing what it calls “responsible AI and production standards” has finalized a set of guidelines for members, Deadline has learned.

The three pillars of the guidelines, Academy officials say, are “creative integrity,” “permissions, licenses, legal and commercial viability”; and “accountability, transparency and sustainability.”

The document resulted from a survey of the organization’s 30,000-plus members. Plans call for rolling them out to the membership following Sunday’s Primetime Emmys.

Generative AI has been the third rail of Hollywood since the arrival of ChatGPT and the labor impasse of 2023, when SAG-AFTRA and the Writers Guild went on strike amid widespread anxiety about AI eliminating industry jobs. Anxiety over copyright protection continues to run high, with Disney, NBCUniversal and Warner Bros. Discovery filing lawsuits against AI image creator Midjourney, one of several legal cases being closely tracked across Hollywood.

The Academy’s task force was led by Christina Lee Storm, governor of the Academy’s Emerging Media Programming peer group, co-chair of the Advocacy Committee, and a member of the Innovation Advisory Committee. She is also a noted producer and co-founder of tech consultancy Playbook PLBK.

Speaking with Deadline this week during the Toronto Film Festival, Storm said the prevailing attitude toward GenAI in the creative community has changed significantly over the past 24 months.

“Two years ago, the Television Academy did an AI Summit [with] a lot of disgruntled folks,” she said, and participants were “not quite understanding the tools, or reading the headlines saying, ‘It’s going to take over everything.’

“This year, we had a lot more people come and be much more open to it, because what we’re seeing – as more and more people are using these tools – is that it’s advancing certain pieces [of the craft], but it’s not advancing everything,” Storm added.

“The need to have a filmmaking language, filmmaking understanding, and all of the [human] traits that are a part of the process – and have always been a part of the process – is crucial.”

According to Lee, the three board-approved ‘Key Principles’ at the heart of the task force’s findings are:

1) Creative integrity for professionals, creators, performers and craftspeople

How does my decision to use GenAI support and respect the work of the artists and collaborators on this project, including writers, performers, directors, producers, and craftspeople? Have they been informed and are properly credited or compensated for their contributions?

2) Permissions, licenses, legal and commercial viability

Is the AI model I’m using trained on ethically sourced, properly licensed and clean data, and not pirated or publicly available content without prior consent or permission from copyright owners? (Storm noted Anthropic’s $1.5 billion settlement, announced Monday, paid out to authors whose work was used without permission.)

3) Accountability, transparency and sustainability

How will I disclose transparency to the production team, distributors, clients, or stakeholders about when, where, and how GenAI was used in the project?

For now, the principles are just recommendations, Storm explained, and the Academy is “not necessarily mandating things per se.” However, the input process involved all 31 of the Academy’s peer groups, a wide spectrum including animators, casting directors, VFX artists and writers.

“The Television Academy recognizes the need for ethical, responsible rules of engagement for the use of Generative AI in television,” Maury McIntyre, president and CEO of the Television Academy, told Deadline. “We commend the Academy’s Innovation Advisory Committee for the development of these Key Considerations and we are committed to educating our members, our colleagues, and our elected representative on how to navigate, adapt and preserve creative integrity as the industry moves forward.”

The news from the TV industry follows a move by the Academy of Motion Picture Arts and Sciences last April to update its rules for Oscar consideration. That decision came after a contentious awards season featuring several AI scandals, including one that dogged Best Picture nominee The Brutalist. AMPAS stated, in somewhat vague terms, that the use of AI and other generative tools would “neither help nor harm the chances of achieving a nomination” going forward.

Storm joined a TIFF AI panel on Monday that also featured Nikola Todorovic, co-founder of Tye Sheridan’s AI, 3D and VFX platform Wonder Dynamics; and Ángel Manuel Soto, the Puerto Rican filmmaker whose credits include DC’s Blue Beetle and HBO Max’s Menudo: Forever Young. Also on the panel was Seungwoo Kim, creator of CJ ENM’s Cat Biggie, a fully GenAI-based 3D animated series; and Don Allen Stevenson III, the creative and AI research lead at Asteria and Moonvalley, which develops workflows for the emerging arena of “ethical” AI in filmmaking.