> ## Content Index > Fetch the complete content index at: https://www.theamericanquorum.com/llms.txt > Use this file to discover other available public pages before exploring further. # Nolan Presses Studios to Defend Film Copyrights Against AI - URL: https://www.theamericanquorum.com/nolan-presses-studios-to-defend-film-copyrights-against-ai/ - Published: 2026-11-08T08:20:51.000Z - Updated: 2026-11-08T08:20:51.000Z - Description: Christopher Nolan says studios must enforce film copyrights as the DGA turns from negotiated AI safeguards to the use of movies as training data. Existing guild protections and copyright law address different parts of that dispute. - Author: News Desk - Tags: Entertainment Christopher Nolan called on Hollywood studios to take responsibility for defending film copyrights against unauthorized artificial-intelligence uses, arguing that directors generally do not own the rights needed to pursue that fight themselves. His comments in a Nov. 7 podcast interview put copyright enforcement alongside workplace protections as a continuing priority for the Directors Guild of America. As [TheWrap reported](https://www.thewrap.com/creative-content/movies/christopher-nolan-previews-dgas-next-phase-of-ai-protections-its-down-to-the-studios/?ref=theamericanquorum.com), the DGA president praised negotiated AI safeguards but said studios must protect the copyrights they hold. He also welcomed language in the closing credits of his film *The Odyssey* stating that it cannot be used as generative-AI training data. Those remarks express a filmmaker and union leader’s position; they do not establish that every disputed training use is unlawful or that a credit notice settles the issue. ## Contract Protections Address Production Work The DGA has already negotiated rules governing AI within covered film and television production. In its June 12 [agreement summary](https://www.dga.org/News/PressReleases/2026/260612%5FDGA%5FBoard%5FRecommends%5FRatification%5Fof%5FNew%5FContract?ref=theamericanquorum.com), the guild said the new contract renewed protections requiring its work to be performed by a person. It also described enhanced provisions establishing directors’ control over AI-generated footage and a new employer-funded skills program. Those provisions concern who performs covered work and who controls footage within production. They address a different question from whether an outside technology developer may copy a completed movie to train a model. Nolan’s latest argument therefore reaches beyond how a studio uses AI on a set or in postproduction: it concerns what happens when films become inputs to another company’s system. The guild announced [member ratification](https://www.dga.org/news/pressreleases/2026/260625%5Fdga%5Fmembership%5Fratifies%5Fnew%5Fcontract%5Fby%5Foverwhelming%5Fmargin?ref=theamericanquorum.com) on June 26, describing the deal as a four-year agreement. Its statement did not publish a numerical voting margin. The agreement is an existing bargaining result, rather than a new contract announced with Saturday’s interview. ## Copyright Ownership Creates a Separate Responsibility [Nolan’s distinction](https://deadline.com/2026/11/christopher-nolan-dga-holding-studios-account-protecting-ai-1237216368/?ref=theamericanquorum.com) turns on the gap between making a film and owning its copyright. In the interview, he said that, with few exceptions, guild members are not the copyright holders of the movies they work on. His demand places responsibility on the companies controlling those rights to challenge uses they consider unauthorized, while the union continues to represent members’ employment and creative interests. He also criticized treating AI as one undifferentiated technology and said litigation and legislation would remain part of the response. The practical implication of that distinction is that a negotiated safeguard cannot answer every dispute. A director’s authority over generated footage, an owner’s permission to copy a film and the legal protection available for a resulting work are separate questions. ## Fair Use Does Not Have a Single AI Answer Nolan expressed confidence that broad fair-use defenses would face problems in court. The U.S. Copyright Office’s May 2025 [training report](https://www.copyright.gov/ai/Copyright-and-Artificial-Intelligence-Part-3-Generative-AI-Training-Report-Pre-Publication-Version.pdf?ref=theamericanquorum.com) offers a more conditional framework: some training uses may qualify as fair use and others may not. The pre-publication report examines differences in purpose, copying, access to works and potential market effects, without deciding individual lawsuits. The office identifies competing concerns. Requiring permission for training can raise costs and practical barriers for developers, while unauthorized copying can undermine creators’ control and markets for their work. Its analysis treats those issues as matters requiring facts about particular systems and uses, rather than accepting either universal permission or a universal prohibition. The office’s [fair-use guidance](https://www.copyright.gov/fair-use/?ref=theamericanquorum.com) likewise explains that courts weigh four statutory factors and assess cases individually. Those factors concern the purpose and character of the use, the nature of the original work, the amount taken and the effect on the potential market. No fixed quantity of copying automatically supplies a safe harbor, and a commercially valuable technology does not itself determine the result. ## Training Rights and Generated Outputs Are Different The distinction also matters after a model produces an image or film sequence. In its January 2025 [copyrightability announcement](https://www.copyright.gov/newsnet/2025/1060.html?ref=theamericanquorum.com), the Copyright Office said AI assistance does not automatically disqualify a larger human-created work from protection. Human authorship remains central, and merely supplying prompts does not provide sufficient control over generated expressive elements. The office identified creative selection, arrangement and modification of generated material as possible sources of human authorship. It did not recommend changing existing law to extend additional protection to AI-generated outputs. For filmmakers, that means the question is what human expression is present in the finished work, rather than simply whether a production used a tool carrying an AI label. That distinction is narrower than a judgment about the artistic quality of the result. That addresses protection for an output, not permission to use someone else’s work in training. A production can contain protectable human contributions while still raising a separate question about source material. Conversely, a dispute over training does not establish that every use of AI in filmmaking eliminates copyright protection. Nolan’s intervention highlights where Hollywood’s next disputes may concentrate: the relationship between negotiated creative control and the enforcement choices of copyright owners. The contract documents establish protections within covered production, while the Copyright Office’s analyses explain why training and output questions require separate treatment. What Saturday’s comments add is a public demand that studios take responsibility for the rights directors say they cannot enforce on their own.