Kathleen Kennedy and Christopher Nolan Bring Hollywood's Secret AI Meetings Into the Open with a Three-Category Copyright Framework

In November 2023, a small group of filmmakers, studio executives, and technologists gathered at artist Refik Anadol's Los Angeles studio. Kathleen Kennedy and Susan Ruskin called the meeting. Nobody announced it. Over the next three years, similar gatherings continued privately at the American Film Institute, slowly expanding in scope. On August 15, 2026, that group held its first public event, drawing 120 leaders from studios, streaming platforms, labor unions, government, and higher education. The framework they released is the first written document Hollywood's own executives have produced to define what AI generated work is, and what copyright protection it can and cannot carry.

The Ankler reported the summit details on August 20. Variety covered the framework's copyright implications the same week. AFI and Google hosted the August 15 event, with Mira Lane, Google's VP of Technology and Society, joining Kennedy and Ruskin on stage.

Three Years of Private Meetings

Kennedy was still president of Lucasfilm when the meetings started. Ruskin, dean and executive vice president of the American Film Institute, organized the meetings alongside Kennedy from the beginning. Their stated goal was to create a space where industry leaders could work through the AI question without the pressure of public position taking. The meetings stayed private long enough for the group to move past the early debates and work toward a shared vocabulary.

Christopher Nolan joined the effort in his capacity as DGA president and chair of the guild's AI committee, a position he has held since his election in September 2025. His DGA leadership has centered on AI governance through negotiation and contractual guardrails, and the AFI group gave him a cross industry forum to develop the same kind of precision he applied to the guild's contract language.

The expansion from a private group to a 120-person summit reflected a deliberate choice. The framework the group had developed needed public testing before it could function as an industry reference. A private document seen by thirty people is a conversation. A document presented to studios, streamers, labor organizations, government officials, and educators in the same room becomes something with institutional weight.

The Three Categories

The framework proposes three categories for AI assisted creative work. Each carries different implications for copyright protection.

Utility Techniques and Embedded AI covers uses of AI that assist production without replacing the human creative act. This includes tools that assist color grading, noise reduction, and subtitle generation, where a human operator makes the meaningful decisions. Work produced this way is copyrightable. The AI functions as an advanced tool.

Human Generative Workflows (HGW) is the category the group defined. It covers creative work where a human uses generative AI as part of a deliberate process. The filmmaker iterates on image or video outputs, selects among generations, directs the tool with specific intent, and assembles results into a work that reflects human authorship. Work produced this way is copyrightable, the framework argues, because the human's sequence of decisions constitutes authorship even though the generation is machine assisted. This is the category where most commercial AI filmmaking currently sits.

Machine Generative (MG) covers output produced by AI with minimal human direction. A prompt entered and a video accepted with no meaningful iteration or selection does not produce copyrightable work under the framework's argument. The output belongs to the machine, not to a human author. This category addresses the concern that studios or individuals might claim intellectual property protection over low effort AI content.

That structure maps directly to the ongoing legal uncertainty about AI copyright. Courts in the United States have held that AI output without human authorship is not protected. The HGW category is the framework's attempt to define what sufficient human authorship looks like in a production context. By naming HGW specifically, the group gives filmmakers, studios, and lawyers a vocabulary for distinguishing protected work from unprotected output.

Who Attended and What They Worked On

The 120 participants at the August 15 summit included studio development executives, streaming platform representatives, SAG-AFTRA and DGA members, officials from federal and state government, and faculty from film and arts programs. Kennedy described the format as "active participation" rather than a traditional conference. Attendees worked in four organized groups rather than listening to panels.

The four working groups addressed: labor, new roles and compensation; production economics; copyright, authorship and permission; and classification, standards and disclosure. Those four areas map to the practical questions studios, unions, and regulators are actually trying to answer. What jobs change and how are workers compensated? What does AI do to production budgets and timelines? Who owns what, and what can be trained on what? How do audiences and buyers know what a film contains?

The Copyright Question the Framework Addresses

The gap the framework is trying to close is real. AI generation has outpaced both legal precedent and industry practice. The U.S. Copyright Office has repeatedly held that AI generated images and text without human authorship are not protectable. But the line between sufficient human authorship and insufficient human authorship has not been drawn with the specificity that a studio lawyer, a union negotiator, or a filmmaker can act on.

The HGW category is an attempt to draw that line. If a filmmaker uses AI video generation with specific, iterated direction, selecting outputs, adjusting prompts over many generations, and assembling the result into a finished work, that filmmaker has done something qualitatively different from typing one prompt and accepting the first result. The framework argues the first process produces copyrightable work. The second does not. The distinction rests on the density of human decisions made in the process, not on whether AI was involved at all.

That distinction matters for studios acquiring AI assisted films, for investors funding AI production companies, for platforms deciding what to license, and for filmmakers deciding how to document their process. The Creators Coalition on AI, whose 500 members include Oscar-winning filmmakers, has pushed for similar clarity from the legislative side. The AFI framework approaches the same problem from inside the industry rather than from a lobbying position.

What Comes Next

The summit's organizers have described August 15 as a beginning. A second convening is already in planning. The framework document is presented as a starting point for industry discussion, not a final standard. The group is inviting feedback from the creative and technical communities that were not in the room.

The Machine Generative category will be the most contested part of the framework. Studios that have built AI content pipelines have a commercial interest in arguing that their AI generated output carries copyright protection. The framework's position directly challenges that interest. How studios, platforms, and courts respond to this framework will determine whether it becomes an actual industry standard or remains a conversation document.

For independent filmmakers, the Human Generative Workflows category offers a practical anchor. A framework that names their process and argues for its legitimacy in copyright terms is useful regardless of whether courts formally adopt it. It is vocabulary filmmakers can use when their work is challenged.

The group's stated goal was to move beyond theory and fear, establish a shared vocabulary, and keep emerging technology in service of intentional creative decisions directed by artists. Whether a voluntary framework with no enforcement mechanism can achieve that depends on how much institutional weight the studios, guilds, and platforms represented in the room choose to give it.

AI video generation tools designed for deliberate, iterated creative workflows are available at AI FILMS Studio.

---

Sources

The Ankler | Variety | AFI