“The new guidelines around how studios can leverage generative AI will shape film and television writing in the years to come.”
Last week, the Writers Guild of America (WGA) reached a tentative three-year deal to resolve a writer’s strike following a labor dispute with the Alliance of Motion Picture and Television Producers (AMPTP). The deal was reached on September 27, 2023, after a 148-day strike, which was the second-longest in the union’s history. According to USA Today, the WGA’s leadership board has lifted the restraining order barring writers from returning to work, and its members will vote to officially ratify the agreement between October 2 and October 9. In a storyline that at one time would have been considered science fiction, a major point of contention between writers and producers was the use of artificial intelligence in the screenwriting process.
Artificial intelligence refers to a system or technology that emulates human intelligence and cognitive functions such as problem-solving and learning. Generative artificial intelligence is a specific type of AI that can produce various types of content, such as text, imagery, and audio, and was a key issue in the Writer’s Strike. This article discusses how copyright law may or may not apply to AI-generated scripts and other creative works, the most significant case law on this topic so far, and how the conversation between copyright law and AI-generated creative works might evolve.
How We Got Here
The bedrock of United States copyright law is found in Article 1, Section 8 of the Constitution, which promotes “the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.” Specifics of U.S. copyright law are currently embodied in Title 17 of the United States Code, the Copyright Act of 1976 (effective January 1, 1978). Section 102(A) of the Copyright Act goes into detail about what is eligible for copyright protection. A work must be (1) an original work of authorship and (2) fixed in a tangible medium of expression. Additionally, the work must possess “a modicum of creativity.” Feist Publ’ns, Inc. v Rural Tel., Serv. Co., 499 U.S. 340, 342.
The debate around whether the use of emerging technology can result in authorship is nothing new. In 1884, the Supreme Court posed a similar question to that being addressed by today’s copyright lawyers in Burrow-Giles Lithographic Co. v. Sarony. A photographer named Napoleon Sarony had captured a photograph of Oscar Wilde entitled “Oscar Wilde No. 18.” Burrow-Giles Lithographic Company reproduced this photo without Sarony’s permission, and the photographer brought suit, arguing that this was a violation of copyright law. Considering that photography was a new medium, there was a natural question of whether it fit within the traditional definition of authorship. The decision in this case hinged on the fact that Sarony did more than just push a button on a camera. To the contrary, Sarony’s work was a “graceful picture” that was a result of his “original mental conception.” He posed Oscar Wilde in front of the camera, selected and arranged his costume and accessories, and adjusted the lighting to create his desired result. Burrow-Giles Lithographic Co. v. Sarony, 111 U.S. 53, 60. Sarony’s involvement in creating the photo ultimately allowed the Court to consider “Oscar Wilde No. 18” as a work of original authorship meriting constitutional protection.
In recent years, courts have found that in addition to being an original work of authorship, a work must have a human author to be eligible for copyright protection. In Naruto v. Slater, a photographer named David Slater left his camera unattended when a monkey named Naruto used it to take a series of photos of himself. These photos went viral, and one photo became known as the “Monkey Selfie.” People for the Ethical Treatment of Animals (PETA) was a party to the case and argued that authorship under the Copyright Act was broad enough to allow for the monkey to be considered the author of the photograph. However, the Ninth Circuit found that the language in the Act, such as “children,” “grandchildren,” “legitimate,” “widow,” and “widower” all implied humanity and necessarily excluded animals. Therefore, Naruto lacked standing under the Copyright Act and the case was dismissed. Naruto v. Slater, 888 F.3d 418, 426 (9th Cir. 2018).
Section 306 of the Copyright Compendium reflects this same understanding. The Compendium of U.S. Copyright Office Practices is a manual produced by the United States Copyright Office, intended for use primarily by the Copyright Office staff as a general guide to policies and procedures such as registration, deposit, and recordation. The U.S. Copyright Office will only register an original work of authorship if the work was created by a human being. The Compendium explicitly refuses to register works produced by nature, animals, plants, and divine or supernatural beings. Additionally, the Copyright Office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author.
In March 2023, the Copyright Office issued new registration guidance: applicants must disclose the inclusion of AI-generated content in works submitted for registration. Whether or not a copyright is awarded depends on whether the work is one of human authorship using technology as an assisting instrument (similar to Sarony’s use of the camera in Burrow-Giles), or if the expression of the work was generated by the machine. Only the former is eligible for copyright protection.
On August 18, 2023, the United States District Court for the District of Columbia ruled that content generated without human involvement is ineligible for copyright protection. The case in question was Thaler v. Perlmutter. The plaintiff, Steven Thaler, owned an AI-powered computer system called the “Creativity Machine,” which generated a piece of artwork entitled “A Recent Entrance to Paradise.” Thaler contested the Copyright Office’s human authorship requirement by arguing that AI should be acknowledged as an author where it otherwise meets authorship criteria, with the copyright ownership vesting in the AI’s owner. Thaler v. Perlmutter, 2023 U.S. Dist. LEXIS 145823, *3-4. While the Ninth Circuit acknowledged that copyright is designed to adapt to the times, human creativity is a foundational requirement for copyright ownership. The court referenced Burrow-Giles, reiterating that “Oscar Wilde No. 18” was awarded copyright protection due to the fact that Sarony “conceived of and designed the image and then used the camera to capture the image,” and Sarony’s orchestration of the photograph was enough to merit the authorship required for copyright ownership Id. at 12. Because the Creativity Machine was solely responsible for “A Recent Entrance to Paradise,” Thaler’s claim was incompatible with the precedent set by Burrow-Giles. Id. at 20. This precedent, that authorship is synonymous with human creation, has remained consistent in copyright law even as the discipline has evolved. Therefore, Thaler was not eligible for copyright ownership of the work. Id. at 20.
Burrow-Giles, Naruto, and Thaler illustrate that there is no legal precedent allowing for authorship of works produced solely by generative artificial intelligence as it relates to copyright registration. As this article has discussed, human authorship is necessary for copyright protection. Copyright law has grappled with what constitutes authorship since the nineteenth century, and the recent Writer’s Strike is the latest example of how emerging technology complicates this issue. In an article for Wired, Will Bedingfield points out that current AI is quickly encroaching on writers’ livelihoods: it can score in the 93rd percentile on SAT reading exams, and can already produce stories and poems. Much of the strike was the WGA attempting to get ahead of this technology before it becomes a more capable version of itself.
AI in Hollywood
As described in The Ringer, the WGA and AMPTP held opposing views as to how the technology can be used in the writing process. Prior to reaching an agreement, the WGA was willing to allow writers to use AI but drew the line at having the technology affect writing credits. Additionally, WGA leadership wanted the following guardrails around the technology: AI could not write or rewrite literary material, AI could not be considered “source material,” and scripts written under the WGA contract could not be used to train AI (a process known as “scraping”). The AMPTP initially rejected these restrictions and instead offered annual meetings to discuss advancements in technology.
In a piece in Mashable, Sam Haysom details the WGA’s concern that studios would use AI as a starting point for scriptwriting and then underpay writers to simply adapt the material. Rewriting a script is not as lucrative an opportunity for a screenwriter as providing a studio with an original draft. As a result, the WGA feared that the technology would be used to avoid paying writers what they would otherwise be entitled to under WGA guidelines. In a scenario where a writer adapts an AI-generated script, a studio is fulfilling the human authorship requirement for copyright protection while simultaneously limiting the writer’s intellectual property rights to the work.
According to The Hollywood Reporter, in the tentative deal reached between the WGA and AMPTP, the WGA secured protections against AI creating or rewriting original material or being used as a “starting point” for writers to revise. Writers can use the technology if their company allows for it. However, the use of generative AI cannot be required by an employer, and companies must disclose if they provide their writers with AI-generated material during the writing process. Additionally, the WGA reserved the right to assert that the usage of a writer’s material to train AI is prohibited by their agreement or other law. The tentative agreement is intended to prevent a worst-case scenario for a television writer in which a producer could request that the writer edit a script generated by AI, and the writer would not be able to profit as much from his or her work as a result.
Like it or Not, the Future Includes AI
The timeliness of this issue shines a spotlight on the relationship between copyright ownership and authorship throughout the creative industries. This being considered, what are the future implications of AI in the creative industries?
A lawsuit to watch as the new agreement is ratified involves comedian and actress Sarah Silverman. Silverman sued artificial intelligence platforms OpenAI and Meta alleging that neither company sought her permission to use her 2010 autobiography to train their systems. Both platforms use existing works to learn language. Silverman argues that failing to seek her approval is a violation of intellectual property law. In an article in Vulture, a member of Silverman’s legal team stated that AI systems “are built entirely and exclusively on the work of human creators” and “take in the training material to try to emulate it.” As a result, artificial intelligence allows for human ingenuity, while sidestepping giving credit to the creator of the original work. While Silverman’s team makes a fascinating case, some expect an uphill battle: in 2016, the Supreme Court ruled in favor of Google Books when it rejected a claim that the platform’s summarization of texts and making excerpts accessible to users was a violation of copyright law.
Major industry players, in fact, show no signs of backing away from artificial intelligence. According to CBS News, Netflix was building out its AI department in the midst of the Writer’s Strike. Controversially, an AI job that paid up to $900,000 pitched the opportunity to help the streaming service “create great content.” Per an article on TheStreet.com, Netflix later adjusted the wording in the job description to reflect the technology’s usage in “powering innovation, from personalization for members to optimizing our payment processing and other revenue-focused initiatives.” The same article discusses how Disney is also seeking to hire senior machine learning engineers who are responsible for “creating AI-enabled solutions for Disney+, Star+, and ESPN+.” The new guidelines around how studios can leverage generative AI will shape film and television writing in the years to come. The film and television industry are currently at a crossroads of legal ambiguity and emerging technology, and it will be fascinating to see how these issues are resolved.
Image Source: Deposit Photos
Image ID: 668326526