Not Another Article About Copyright and Generative AI

By Katrina Geddes (Cornell Tech & NYU)

Copyright has been getting a lot of attention in the press, and for good reason. Twenty-eight lawsuits have now been filed against AI vendors by artists, authors, news publishers, and record labels for copyright infringement associated with the inputs and outputs of generative AI models.[1] Everyone is eager to share their views on whether generative models are protected by fair use.[2]

As exciting as this speculation is, it has overlooked other areas of intellectual property law where AI vendors may be equally (if not more) liable. Trademark law, for example, intersects with generative AI in at least three different ways. First, AI outputs have been accused of trademark infringement,[3] and dilution.[4] Getty Images’ complaint against Stability AI alleges that some of the model’s outputs contain a modified version of the Getty Images watermark, which creates a false association with Getty Images in the minds of consumers, and dilutes the quality of its marks, especially where they appear on offensive or low-quality AI outputs. Similarly, The New York Times’ complaint against OpenAI alleges that the unauthorized use of the Times’ trademarks in low-quality and inaccurate ChatGPT outputs dilutes the quality of those marks. It may be difficult, however, for plaintiffs to prove that the use of marks in AI-generated outputs constitutes commercial use, and that there is a meaningful likelihood of confusion, especially if the marks are distorted.[5]

Second, companies using AI models to generate logos or brand designs may inadvertently generate marks that resemble the model’s training data. Registration may be refused if the AI-generated mark is confusingly similar to a registered trademark for related goods or services.

Third, visual artists have alleged that AI-generated outputs infringe their trade dress when they replicate the distinctive aesthetic of their works. The plaintiffs in Andersen v. Stability AI claim that Midjourney controls, and profits from, trade dress imitations by encouraging users to generate outputs that reproduce the recurring visual elements and artistic techniques that characterize their work. Trade dress may offer visual artists some protection for their distinctive style, which is generally not protected under copyright law (although see the DLI’s own Ben Sobel for a nuanced treatment of this issue).[6]

Generative AI also raises a host of issues concerning the right of publicity. Generally speaking, the right of publicity gives an individual control over how their likeness is used in commerce. So, if a famous actress declines to lend her voice to an AI-powered chatbot assistant, the AI firm can’t just hire an actor to imitate her voice.[7] (Ford tried this with Bette Midler in the 80’s and it didn’t work out so well for them.) Similarly, if someone trains an AI model to simulate the vocals of two famous artists, those artists can sue for violation of their publicity rights.[8]

Publicity rights are governed by a patchwork of state laws with idiosyncratic requirements, which creates confusion and information costs for litigants. Pursuing bad actors through multiple state court systems is very expensive. There are increasing calls for a federal right of publicity,[9] and bills to this effect have been proposed.[10] Although the right of publicity emerged from the right to privacy (and a desire to protect individuals from dignitarian harm), it has evolved into a property-like interest that encompasses almost any unauthorized use of an individual’s name or likeness. Today, it’s closer to a property right than a privacy right, and it’s trademark-adjacent in the sense that it’s designed to prevent false and misleading representations of endorsement.[11]

Many public figures, including Pope Francis, have been the subject of deepfake imagery.[12] Could the Pope sue the creators of those deepfakes for violating his publicity rights? Probably not. First, those deepfakes were created for non-commercial purposes, and many state publicity laws have a commerciality requirement. Second, the deepfakes would likely be protected by the First Amendment as a “transformative” expressive use. The danger with publicity rights is that individuals will try to suppress any unflattering portrayals of themselves, at the expense of valuable expression, like parody or satire. Accordingly, courts try to resolve the tension between free speech and publicity rights by asking whether the secondary work “adds significant creative elements” that take it beyond “a mere celebrity likeness or imitation.” (Although there is some question about whether this transformativeness test will change after the Supreme Court’s decision in Warhol).

The AI-generated song, “Heart On My Sleeve,” showcases the complex intersection of intellectual property rights implicated by generative AI.[13] The creator (a songwriter operating under the pseudonym Ghostwriter977) explained that he imitated the vocals of Drake and The Weeknd (using so-called “AI vocal filters”) in order to highlight the undervalued contributions of songwriters.[14] The song’s resonance with fans, despite no involvement by recording artists, demonstrates the value that songwriters bring to creative projects.

To create “Heart On My Sleeve,” Ghostwriter977 presumably trained an AI model on copyrighted songs by Drake and The Weeknd in order to generate sound-alike vocals. The creation of these training copies may attract copyright liability if courts ultimately find that unlicensed training is not fair use. The AI-generated output may also attract liability if Drake and The Weeknd can show that it’s substantially similar to their works. However, this may be hard to prove. Ghostwriter wrote his own melody and lyrics, and copyright protection for sound recordings does not extend to deliberate imitations.[15] The musicians’ strongest claim is that the simulation of their voices represents a violation of their publicity rights under state law. However, there is no DMCA takedown remedy for publicity rights, and filing a state claim is a slower process. In order to have the song removed, Universal Music Group reportedly had to reference the unauthorized sampling of a producer tag.[16]

As generative AI raises public awareness of publicity rights, there are signs that AI vendors may be more cautious about licensing the use of celebrity voices for their AI products in the future.[17]

[1] https://chatgptiseatingtheworld.com/2024/08/07/status-of-all-28-copyright-lawsuits-plus-1-new-state-law-suit-v-ai-aug-7-2024-1st-case-involving-youtube-videos-but-no-copyright-claim/

[2] https://twitter.com/Dan_Jeffries1/status/1740303405254377808.

[3] 15 U.S. Code §1114.

[4] 15 U.S. Code § 1125.

[5] https://www.theverge.com/2023/2/6/23587393/ai-art-copyright-lawsuit-getty-images-stable-diffusion

[6] Sobel, Benjamin, Elements of Style: Copyright, Similarity, and Generative AI (May 18, 2024). Harvard Journal of Law & Technology, Forthcoming Vol. 38, Cornell Legal Studies Research Paper Forthcoming, Available at SSRN: https://ssrn.com/abstract=4832872. 

[7] https://www.theguardian.com/technology/article/2024/may/27/scarlett-johansson-openai-legal-artificial-intelligence-chatgpt

[8] https://www.nytimes.com/2023/04/19/arts/music/ai-drake-the-weeknd-fake.html

[9] https://rightofpublicityroadmap.com/news_commentary/federal-right-of-publicity-takes-center-stage-in-senate-hearing-on-ai/

[10] https://www.congress.gov/bill/118th-congress/house-bill/6943/text?s=1&r=9

[11] Mark A. Lemley, Privacy, Property, and Publicity, 117 Mich. L. Rev. 1153 (2019).

[12] https://www.theguardian.com/commentisfree/2023/mar/27/pope-coat-ai-image-baby-boomers

[13] https://www.npr.org/2023/04/21/1171032649/ai-music-heart-on-my-sleeve-drake-the-weeknd

[14] https://www.billboard.com/music/pop/ghostwriter-heart-on-my-sleeve-drake-ai-grammy-exclusive-interview-1235434099/

[15] 17 U.S. Code § 114(b).

[16] https://hls.harvard.edu/today/ai-created-a-song-mimicking-the-work-of-drake-and-the-weeknd-what-does-that-mean-for-copyright-law/

[17] https://www.theverge.com/2024/8/3/24212580/meta-ai-voice-assistant-celebrity-voices-judi-dench-awkwafina-keegan-michael-key

Katrina Geddes

Postdoc, Cornell Tech (DLI) & NYU

kgg283@nyu.edu

More Info >

Cornell Tech | 2024

Download PDF >