Innocent Fun or a Copyright Nightmare?

Actors are increasingly seeing videos of themselves, but they don’t remember taking them. This is an issue Hollywood actors, talent agencies, and studios are facing with artificial intelligence (“AI”), specifically OpenAI’s Sora 2. (Wendy Lee and Samantha Masunaga, LA Times). Sora 2 generates realistic videos with synchronized audio from user-entered text prompts. (OpenAI). Sora 2 has technology that can generate recognizable properties or likenesses into deepfakes — either friends, copyrighted characters, or famous likenesses. (Winston Cho, Hollywood Reporter). A deepfake is a video that seems authentic but has been manipulated by AI; sometimes used for disinformation or extortion. (Government Accountability Office, Science, Technology Assessment, and Analytics). In one such video, viewers can see Michael Jackson interacting with Bryan Cranston, of “Breaking Bad.” (Wendy Lee and Samantha Masunaga, LA Times). Thus, the controversy revolves around who owns the copyrighted images and human likenesses that are utilized to create such videos. Id. This post will discuss a brief overview of copyright law; why Sora 2 is alarming Hollywood; how agencies and studios are protecting their clients and intellectual property; and the implications of the rising use of Sora 2 and similar technologies.

Copyright protections are codified in federal law and give the rightsholder the exclusive ownership and control over reproductions, derivatives, distribution, sale, performance, or display of a variety of creative works, including “motion pictures and other audiovisual works.” (17 U.S.C. §§ 102, 106). These protections are automatic upon the creation of the work, though registration is required before a rightsholder can bring a lawsuit for infringement. Valancourt Books, LLC v. Garland, 82 F.4th 1222, 1227 (D.C. Cir. 2023). Under copyright law, entities and individuals interested in using copyrighted material must obtain permission from the rightsholder in advance of use. (Wendy Lee and Samantha Masunaga, LA Times). Considering that an AI platform, like Sora 2, can produce works without asking a rightsholder for permission to use their copyright, systemic technology shifts and advancements, like Sora 2, are changing how works are created, disseminated, and consumed by the public.

The entertainment industry is increasingly concerned with the increased use of AI tools, such as Sora 2. Sora 2 may infringe on a rightsholder's ability to protect their work because Sora 2’s generated content incorporates protected characters and actors' likenesses. In initially creating content, Sora 2 circumvents the copyright use request process and defaults to using copyrighted works without permission unless a rightsholder explicitly forbids the use of the work. (Aaron Moss, Copyright Lately). This cuts against the essence of copyright law and modifies the current regime to an op-out one, requiring the rightsholder to take affirmative action to exclude their protected content. (Jared Perlo, NBC).

Hollywood hasn’t taken OpenAI’s blatant disregard of copyright laws without protest. Influential Studios stated to OpenAI that all their actors and clients would be opting out. (Wendy Lee and Samantha Masunaga, LA Times). The Motion Picture Association (“MPA”), a lobbying group, alongside major talent agencies, have decided that their clients would not be participating in Sora 2. (Winston Cho, Hollywood Reporter). The MPA, along with others, stated that it is OpenAI’s sole responsibility to prevent infringement through its service — in line with copyright principles. (Ted Johnson, Et al., Deadline). Whereas, Disney sent a letter to OpenAI firmly outlining Disney’s position that it does not consent to its copyrighted property being used by Sora 2. (Jaures Yip, CNBC). Disney reiterated the long-standing copyright principle that it has no obligation to opt-out to protect its own property. Id. Some studios have escalated their fight by suing AI firms for copyright infringement, hoping to quell the proliferation of unauthorized copyright use. (Wendy Lee and Samantha Masunaga, LA Times). Industry giant Warner Bros. Discovery reiterated the fixed nature of the traditional principles of copyright protections and expressed outrage at OpenAI’s current practices with Sora 2. (Wendy Lee and Samantha Masunaga, LA Times).

In response to the industry outcry and the ever-increasing pile of lawsuits, OpenAI switched Sora 2 to an opt-in model. (Cody Combs, The National). OpenAI claims they have implemented “more granular controls to rightsholders” to allow rightsholders to oversee how their content is used. (Wendy Lee and Samantha Masunaga, LA Times). Furthermore, OpenAI instilled “guardrails to block the generation of well-known characters.” Id. While Open AI has taken steps to curtail the use of copyrighted content, some rights holders still need to opt out and make manual requests to remove some content. Id.

Even so, OpenAI appears to take the matter seriously and is supporting legislation to protect unauthorized AI-generated content by supporting the NO FAKES Act. (Jaures Yip, CNBC). The NO FAKES Act was introduced in 2024 by a bi-partisan coalition of U.S. Senators for consideration and supported by Hollywood. (Chris Coons, Et al. Senate.gov). The bill aims to protect individuals from unauthorized AI-generated content using their voices or likenesses. Id. The bill is still in the beginning stages, and it is unclear if it will pass because of its overly broad scope. (Trevor Wagner, CCIA). However, OpenAI’s motivations remain unclear, considering the bill is aimed at protecting everyday citizens and not necessarily protected characters. Id. Regardless, Congress and the courts have proven to be historically slow to act in response to emerging technologies, unless they pose an imminent national security threat, threat to public safety, or can be capitalized on. (Allyson Brunette, Thomson Reuters).  

Although AI has many benefits, this new technology also brings unprecedented legal concerns. The proliferation of Sora 2 could shape the future of the entertainment industry and create new legal implications of using AI. (Wendy Lee and Samantha Masunaga, LA Times). Without limits and regulation, new AI developments could enable unchecked, unauthorized access to copyrighted materials and undercut the entertainment industry. (Ted Johnson, Et al., Deadline). Sora 2 offends traditional copyright laws, but OpenAI is seemingly now helping curb infringement on Hollywood’s intellectual property. Regulating these new emerging technologies could prove essential to preserve traditionally vested rights and protect from potential nefarious misuse.