By: Maddie Rudge
Deepfaking Artificial Intelligence has the capability to damage many people: both famous and obscure. For example, deepfake AI images were created showing Taylor Swift doing a variety of obscene activities.[1] People around the world talked about Taylor Swift doing obscene activities, gaining over 40 million views even though these images were obviously fake.[2] The same thing could be done to a local kid in their high school, completely damaging their reputation in their local community. Individuals using AI also have the capability to mimic someone else’s voice, making it say whatever that person wants it to say. For celebrities, this could mean using an actor’s voice in advertisements without getting permission from the actor or compensating them. For a musician, this could mean using their voice to create new music without their permission. In fact, this has already happened. For instance, deepfake AI has been used to create new songs using the voice and styles of both Tupac and Freddie Mercury.[3] For the ordinary person, this could mean doctoring a voice recording making others believe they said things they never said. As AI continues to evolve, strangers on the internet will have difficulty discerning a real video or voice recording from an AI generated image.
To protect our state’s robust music community, Governor Bill Lee introduced the Ensuring Likeness Voice and Image Security (ELVIS) Act, which will update Tennessee’s right of publicity law to protect songwriters, performers, and music industry professionals’ voice from the misuse of AI.[4] Although this bill focuses on musicians and other artists, it will protect all citizens of Tennessee, despite their level of fame.[5] This bill is the first of its kind to be proposed in a state[6], and comes right after lawmakers from the U.S. House of Representatives revealed the No Artificial Intelligence Fake Replicas and Unauthorized Duplications Act (No AI FRAUD Act).[7] The No AI FRAUD Act “aims to establish a framework for protecting one’s voice and likeness on a federal level and lays out First Amendment protections.”[8] If this bill passes, it will be the first federal-level right of publicity.[9]
The ELVIS Act and the No AI FRAUD Act also protects artists, both new and old, from exploitation. If people are able to freely create new songs from artists and bands long past, such as the Beatles or Elvis Presley, it will be that much harder for new artists to break into the business. Also any of these old artists or bands (or their estates) would not be paid for such use, instead they would likely be surprised by it along with the rest of the world. That was the case for Drake and The Weeknd about a year ago.[10] A new single by both of the artists began trending in April 2023, gaining 600,000 streams on Spotify, 15 million views on TikTok, and 300,000 views on YouTube.[11] It was finally removed after complaints because neither artist had actually created the song: it was generative AI.[12] Drake and the Weeknd hadn’t even known it had existed until it went viral and they did not receive any compensation.[13]
The bills also protect the careers of voice actors. There are many actors that we may not be able to identify them by their face, but we know who they are as soon as we hear them speak. Without these protections, companies would be able to use these actors’ voices to promote their own products or create content without compensating the actors. For example, in Roadrunner, a documentary about the late Anthony Bourdain, they used AI to clone Anthony Bourdain’s voice and had it read an email he had written.[14]
While both bills have a lot of support from the music industry[15], there are still some concerns. Critics state that the No AI FRAUD Act is too broad, because it “applies to an incredibly broad amount of digital content: any ‘likeness’ and/or ‘voice replica’ that is created or altered using digital technology, software, an algorithm, etc.”[16] This would mean this legislation would apply to virtually every type of digital content that portrays a human.[17] Similarly, some believe the ELVIS Act is also too broad. For example, if a sound engineer autotunes a singer’s voice to sound a bit like a famous singer, at what point would such a change violate the ELVIS Act?[18]
Still, there may be some causes of action that arise under both the ELVIS Act and the No AI FRAUD Act that were not meant to originally be covered by such legislation. This is also a common problem in groundbreaking legislation such as this. All laws have been challenged in the court system in ways that the original drafters never intended. Even today, we continue to challenge our original laws in unique ways, for example the Bill or Rights.
Although both Acts are not perfect right now, litigation will expose their biggest issues if they are enacted. Then, such issues will be resolved by the court system. If the legislative body does not like those solutions, or finds the court’s solutions to be lacking, then they will propose amendments to the bills. The current priority lies in safeguarding individuals, regardless of their fame, through the use of generative AI. It’s imperative that nobody should have their voice or likeness manipulated to portray actions or utterances they haven’t actually made.
Every person should have control over their own voice and image, no matter what technological advances are made. Tennessee is the first State to create legislation protecting its citizens from deepfaking AI, ahead of the curve in a progressing issue. Other states should follow in Tennessee’s foot steps to create protections against generative AI.
[1] Robert Westermann et al., Students and Experts React to Taylor Swift AI Drama, USC Annenberg Media (Jan. 30, 2024, 6:09 PM), https://www.uscannenbergmedia.com/2024/01/30/students-and-experts-react-to-taylor-swift-ai-drama/.
[2] Id.
[3] Srishreya (Shreya) Arunsaravanakumar, Deepfake Music Sends Ripples Across the Music Industry, Wildcat Trib. (May 25, 2023), https://thewildcattribune.com/17528/ae/deepfake-music-sends-ripples-across-the-music-industry/.
[4] Edward D. Lanquist et al., New Legislation Proposed in Tennessee to Protect Against Deepfakes: ELVIS Act, Baker Donelson (Jan. 17, 2024), https://www.bakerdonelson.com/new-legislation-proposed-in-tennessee-to-protect-against-deepfakes-elvis-act.
[5] Id.
[6] Id.
[7] Kristin Robinson, Tennessee State Bill to Protect Artists’ Voices from AI Impersonation Unveiled By Governor, Billboard (Jan. 10, 2024), https://www.billboard.com/business/legal/tennessee-ai-bill-defend-artists-voice-impersonation-1235579559/.
[8] Kristin Robinson, House Lawmakers Unveil NO AI FRAUD Act in Push for Federal Protections for Voice, Likeness, Billboard (Jan. 10, 2024), https://www.billboard.com/business/legal/no-ai-fraud-act-congress-federal-law-explained-1235578930/.
[9] Nathaniel Bach & Tal Edelstein, U.S. House Proposes ‘No AI FRAUD Act’ to Protect Individuals’ Likeness and Voices, JD Supra (Jan. 17, 2024), https://www.jdsupra.com/legalnews/u-s-house-proposes-no-ai-fraud-act-to-3223759/.
[10] Josh Heier, Is This Now the Biggest Threat to Music Royalties?, Asset Scholar (Jan 22, 2024), https://assetscholar.com/ai-and-music-royalties/.
[11] Id.
[12] Id.
[13] Id. The same thing happened to Bad Bunny and Justin Bieber.
[14] Amanda Hoover, Voice Actors Are Bracing to Compete with Talking AI, Wired (Oct. 3, 2023, 7:00 AM), https://www.wired.com/story/ai-voice-actors-jobs-threat/#:~:text=The%20tech%20doesn’t%20just,that%20would%20damage%20their%20professional.
[15] Audrey Gibbs, ELVIS Act protecting Tennessee Musicians Against AI Abuse Advances in the House, The Tennessean (Feb. 14, 2024, 1:39 PM), https://www.tennessean.com/story/entertainment/music/2024/02/14/elvis-act-protecting-tennessee-musicians-against-ai-abuse-advances/72598673007/.
[16] Corynne McSherry, The NO AI Fraud Act Creates Way More Problems Than It Solves, Elec. Frontier Found., (Jan. 19, 2024), https://www.eff.org/deeplinks/2024/01/no-ai-fraud-act-creates-way-more-problems-it-solves.
[17] Id.
[18] Lanquist et al., supra note 3.