Tom Graham, CEO of generative AI tech company Metaphysic — the company known for popularizing a deepfake TikTok account that spoofs Tom Cruise — underscores the growing debate over ownership of one’s likeness in the age of AI and Deepfakes . He has his AI resemblance to the US Copyright Office. He seeks to demonstrate how copyright laws can be used to protect against the unauthorized creation and distribution of content that duplicates a person’s likeness through AI.
As Hollywood grapples with ethical and legal questions surrounding its use of the technology, the US Copyright Office confirmed in March that most AI-generated works are not copyrighted, but clarified that they are eligible for protection in certain cases.
Speaking at CES in January, SAG-AFTRA National Executive Director and Chief Negotiator Duncan Crabtree-Ireland stressed the importance of addressing this issue. He reported that the guild is working on policies aimed at creating opportunities – not taking them away – while protecting its members from exploitation of their likenesses.
In the United States, copyright laws do not protect exclusively machine-created works. Several courts have ruled that copyrights can only be granted on works created by humans. An application containing AI-generated material can only support a claim if it satisfies the requirement that a human contributed to the majority of the work, for example if a person “selected” the piece in a “sufficiently creative way” or arranged” so that the resulting work is an original work of authorship.”
In his copyright application, Graham lists himself as the author of the work. This is a crucial distinction from a copyright filing by Stephen Thaler, the CEO of neural network company Imagine Engines, who credited an AI system as the creator of a work of art called A Recent Entrance to Paradise. The copyright office denied registration, saying the work “lacks the human authorship necessary to support a copyright claim” and that “the connection between the human mind and creative expression” is a critical element of protection. Last year, Thaler sued the office to challenge the rejection.
Contrary to the filing from Thaler and others seeking protection for AI-assisted works, Graham emphasizes his contribution in creating the video he seeks copyright for. He says he put together the dataset used to train the model and added “other man-made work” to help the Metaphysic AI tool successfully produce a realistic image of itself. There was also assembly and editing work in the process, he notes.
“All of these steps are very deliberate, human-driven activities, which I think are indistinguishable from someone using Photoshop or VFX tools,” says Graham. “We use AI, but there is a great deal of man-made work and human effort. It is fundamentally me from top to bottom.” Perhaps most recognized for powering the @DeepTomCruise account develops Metaphysic generative AI tools and services for talent, including through a partnership with CAA.
The viability of a copyright will largely depend on how the AI tool works and how it was used to create the final work. If an AI technology only receives a prompt and produces a work, for example, it is ineligible for protection because the traditional elements of human authorship are determined and performed by the technology.
Copyright protection standards are still up in the air. Suzanne Hengl, an intellectual property attorney who does not represent Metaphysic, is skeptical that Graham’s application will meet the agency’s current human authorship standards. “As far as there was extensive training to get the AI tool working and creating an output, you can’t get over it,” she says.
Law firm partner Baker Botts notes there may be protection if work was completed by a human after the video was made. She points to copyright office guidelines that an artist using Photoshop to edit an image qualifies for protection of the altered image.
While Graham’s video may not be copyrighted, the CEO’s filing highlights the lack of protection against the unauthorized distribution of videos and photos in which an individual’s likeness is duplicated through AI. There are no copyright laws specifically designed to counter the use of deepfakes. In fact, they allow them in some cases.
Several cases of deepfakes in Hollywood fall under the “fair use” exception to copyright infringement. The doctrine allows unauthorized use of copyrighted works, such as commentary, criticism, and news coverage, under certain circumstances. One of the four factors that determine eligibility is the purpose and nature of the use, which extends protections for transformative works when the purpose of the copyrighted work is changed to create something with a new meaning or message. to create.
The fair use exception to copyright law is why Kendrick Lamar had legal footing when he used deepfakes to turn into Will Smith, Jussie Smollett and Kobe Bryant in the music video “The Heart Part 5”. They did not agree to star in the video, but the deepfakes would likely be considered transformative use to explore black representation in Hollywood, in addition to the music video’s other themes. They could also have been categorized as parodies, which is also protected as transformative use under copyright laws.
One of the reasons Graham says he signed up to register the AI-created video of himself is to pave a path for people to protect their identities, performances and brands from unauthorized content created with using AI tools, such as deepfake pornography. Under the Digital Millennium Copyright Act, a copyright owner can quickly remove material that infringes its copyright.
“I want to underline the value of using the DMCA as a means of removing an infringing video within 24 hours,” said Jordan Manekin, an attorney for Metaphysic. “That’s a much more readily available remedy for most people than just going to court and litigating for several years.”
But even if protection is granted, Hengl says it’s unlikely Graham will be able to use his copyright to take down AI-created photos and videos himself. “In terms of being able to use copyright law as a way to enforce rights related to a likeness or to enforce privacy rights, it’s really trying to fit a square peg into a round hole,” she argues . “You can only enforce protection against works that are essentially similar. There must be more than that person’s image – their clothing, expression on their face, the posture their body is in, the background they are in – that is similar.
Hengl notes that the traditional course of action against deepfakes and similar technology has been to pursue claims for violations of rights of publicity and privacy.
As SAG-AFTRA’s Crabetree-Ireland summarized at CES, “As this technology democratizes and becomes more accessible and available, it will become a broader public policy issue.”