Sure, everyone hates record labels \u2014 but the AI industry has figured out how to make them look like heroes. So that\u2019s at least one very impressive accomplishment for AI.<\/p>\n
AI is cutting a swath across a number of creative industries \u2014 with AI-generated book covers<\/a>, the Chicago Sun-Times publishing an AI-generated list of books that don\u2019t exist,<\/a> and AI-generated stories at CNET under real authors\u2019 bylines.<\/a> The music industry is no exception. But while many of these fields are mired in questions about whether AI models are illegally trained on pirated data, the music industry is coming at the issue from a position of unusual strength: the benefits of years of case law backing copyright protections, a regimented licensing system, and a handful of powerful companies that control the industry. Record labels have chosen to fight several AI companies on copyright law, and they have a strong hand to play.<\/p>\n Historically, whatever the tech industry inflicts on the music industry will eventually happen to every other creative industry, too. If that\u2019s true here, then all the AI companies that ganked copyrighted material are in a lot of trouble.<\/p>\n Can home prompting kill music careers<\/strong>?<\/p>\n There are some positive things AI music startups can accomplish \u2014 like reducing barriers for musicians to record themselves. Take the artist D4vd, who recorded his breakout hit \u201cRomantic Homicide\u201d in his sister\u2019s closet using BandLab<\/a>, an app for making music without a studio that includes some AI features. (D4vd began creating music to soundtrack his Fortnite YouTube montages<\/a> without getting a copyright strike for using existing work.) The point of BandLab is giving more musicians around the world the opportunity to record music, send it into the world, and maybe get paid for their work, says Kuok Meng Ru, the CEO of the app\u2019s parent company. AI tools can supercharge that, he says.<\/p>\n That use, however, isn\u2019t exactly what big-time AI companies like Suno and Udio have in mind. Suno declined to comment for this story. Udio did not respond to a request for comment.<\/p>\n Suno and Udio are designed to let music consumers generate new songs with a few words. Users type in, say, \u201cPrompt: bossa nova song using a wide range of percussion and a horn section about a cat, active, energetic, uptempo, chaotic\u201d and get a song, wholesale, without even writing their own lyrics<\/a>. The idea that most listeners will do this regularly seems unlikely \u2014 making music is more work than just listening to it, even with text prompts \u2014 as does the idea that AI will replace people\u2019s favorite human artists. (Also, the music is pretty bad.) <\/p>\n \u201cAI flooded the market with it.\u201d<\/p>\n A lot of listening is passive consumption, like a person putting on a playlist while doing the dishes or studying, or a business piping background tunes to customers. That background music is up for grabs \u2014 not by consumers, but by spammers using these tools. They\u2019re already generating consumer-facing slop and putting it on Spotify<\/a>, effectively crowding out real artists.<\/p>\n That seems to be the major use case for these apps. Generating a two-minute song on Udio costs a minimum of eight credits; free users get around 400 credits monthly; for $10 a month, you\u2019ll get 1200, the equivalent of, at most, 150 songs. Spotify Premium individual costs $12 a month and gets you just about everything ever recorded, plus audiobooks. Also, it takes many, many fewer clicks to listen to Spotify than it does to generate your own songs \u2014 so if you\u2019re looking for something to listen to while you cook, Spotify is just easier.<\/p>\n But the math there changes if you\u2019re looking for background music for your YouTube videos \u2014 or anything else that\u2019s meant to be listened to publicly. That means AI music threatens people who support themselves by making incidental music for advertisements, or recording \u201cperfect fit content\u201d for Spotify<\/a>, or other, less-glamorous work. Taylor Swift\u2019s career isn\u2019t endangered by AI music \u2014 but the real people who make the background music for Chill Beats to Study To, or the hold music you hear on the phone, are.<\/p>\n \u201cI wouldn\u2019t want to be [new-age musician] Steven Halpern and have my future career based on meditation music,\u201d says David Hughes, who served as CTO for the Recording Industry Association of America (RIAA) for 15 years. He now works as a tech consultant for the music industry at Hughes Strategic. \u201cAI flooded the market with it. There\u2019s no business making it anymore.\u201d<\/p>\n As in other creative industries, AI music tools are poised to hollow out the workaday middle of the market. Even new engineering tools have their downsides. Jimmy Iovine, who eventually founded Interscope Records and Beats Electronics, started his career as an audio engineer before making his name by producing Patti Smith\u2019s Easter. This is kind of like starting in the mail room and becoming the CEO; if more of the engineering work is done by AI, that removes career paths. The next Jimmy Iovine might not get his start, Hughes says. \u201cHow does anyone apprentice?\u201d<\/p>\n And it\u2019s (possibly) illegal<\/strong><\/p>\n About a year ago, the major labels brought suit against Suno and Udio<\/a>. The fight is about training data; the labels say the companies stole copyrighted work and violated copyright law by using it to build their models. Suno has effectively admitted it trained its AI song generator on copyrighted work<\/a> in documents filed in court; so has Udio<\/a>. They\u2019re saying it was fair use, a legal framework under which copyrighted work can be used to create new work. <\/p>\n Virtually every creative industry is in some kind of similar fight with AI companies. A group of authors is suing Meta, Microsoft, and Bloomberg<\/a> for allegedly training on their books. The New York Times is suing Microsoft and OpenAI<\/a>. Visual artists have sued Stable Diffusion and Midjourney<\/a>; Getty Images is also suing Stable Diffusion<\/a>; Disney and Universal are suing Midjourney<\/a>. Even Reddit is suing Anthropic. Training data is at issue in all the suits.<\/p>\n \u201cThou shalt not steal.\u201d<\/p>\n So far, the legal takes on AI have been contradictory, and at times, baffling. There doesn\u2019t seem to be a consistent through line, so it\u2019s hard to know where the law will ultimately end up. Still, music has its own legal history that comes to bear \u2014 from unauthorized sampling. That may mean it\u2019s entitled to stronger protections.<\/p>\n In Bridgeport Music v. Dimension Films, a case about NWA\u2019s sample of Funkadelic\u2019s \u201cGet Off Your Ass and Jam,\u201d the US Court of Appeals ruled that the uncompensated sampling was in violation of copyright law. In the decision, the court found that only the copyright owner could duplicate the work \u2014 so all sampling requires a license. Some other courts have rejected that ruling, but it remains influential. There\u2019s also Grand Upright Music v. Warner Bros. Records, in which the US Southern District of New York ruled that Biz Markie\u2019s sample of Gilbert O\u2019Sullivan\u2019s \u201cAlone Again (Naturally)\u201d was copyright infringement. The written opinion in the case begins, \u201cThou shalt not steal.\u201d<\/p>\n \u201cSome of the sampling cases have suggested that sound recordings might be entitled to stronger protections than other copyrighted works,\u201d says James Grimmelmann, a professor at Cornell Law School. Those protections may extend beyond sampling to generative AI, especially if the AI outputs too closely resemble copyrighted work. \u201cFrom that perspective, music becomes kind of untouchable. You just can\u2019t do this kind of work on it.\u201d <\/p>\n Music is also complicated \u2014 since performances are bound up in rights of publicity. In the case of the fake Drake track,<\/a> the soundalike may violate Drake\u2019s right to publicity. Artists such as Tom Waits and Bette Midler have won suits<\/a> against more mundane human soundalikes. Proving that someone meant to violate Drake\u2019s right to publicity might be even more straightforward if the lawsuit contains the prompt.<\/p>\n This may be an easier case for music companies to make<\/p>\n As in other AI fair use cases, one of the key questions is whether a derivative work, such as \u201cBBL Drizzy,\u201d is intended to replace or disrupt a market for an original one. In 2023, the Supreme Court ruled that Lynn Goldsmith\u2019s copyright had been infringed<\/a> on by Andy Warhol when he screenprinted one of her photos of Prince. One of the key factors was that Vanity Fair had licensed Warhol\u2019s work instead of Goldsmith\u2019s \u2014 and she received no credit or payment.<\/p>\n In May, Register of Copyrights Shira Perlmutter released a pre-publication report<\/a> that found that AI training in general was not necessarily fair use. In the report, one of the factors considered was whether an AI product supplanted the use of the original. \u201cThe use of pirated collections of copyrighted works to build a training library, or the distribution of such a library to the public, would harm the market for access to those works,\u201d the report said. \u201cAnd where training enables a model to output verbatim or substantially similar copies of the works trained on, and those copies are readily accessible by end users, they can substitute for sales of those works.\u201d<\/p>\n This may be an easier case for music companies to make than, let\u2019s say, ad writers. (What copywriter wants to admit they\u2019re so uncreative they can be replaced by a machine, first of all?) Not only are there fewer of them, which allows them to easily negotiate as a bloc, it\u2019s simple enough to point to the output of AI music singing Jason Derulo\u2019s name<\/a>, or mimicking \u201cGreat Balls of Fire.\u201d That\u2019s pretty clear-cut.<\/p>\n Another crucial factor \u2014 one that matters particularly to the music industry \u2014 was lost licensing opportunities. If copyrighted works are being licensed as AI training data, doing a free-for-all snatch and grab robs rights holders of their ability to participate in that market, the report notes. \u201cThe copying of expressive works from pirate sources in order to generate unrestricted content that competes in the marketplace, when licensing is reasonably available, is unlikely to qualify as fair use,\u201d the report says.<\/p>\n The RIAA alleges illegal copying on the front end and infringing outputs on the back end<\/p>\n Recently, Anthropic got a ruling in a copyright case that differs from this analysis. According to Judge William Alsup<\/a> of the Northern District of California, using books for training data is fair play<\/a> \u2014 with two big caveats. First, any inputs must be legally acquired, and second, the outputs must be non-infringing. Since Anthropic pirated millions of books, that still leaves the door open for massive damages, even if using the books to train isn\u2019t wrong. <\/p>\n When it comes to the Suno and Udio suits, the RIAA alleges illegal copying on the front end and infringing outputs on the back end, Grimmelman says. Suno and Udio can introduce evidence to rebut those allegations, but the ruling isn\u2019t ideal to knock down the RIAA\u2019s suit. It\u2019s also not clear Suno can rebut those allegations. \u201cSuno\u2019s training data includes essentially all music files of reasonable quality that are accessible on the open Internet, abiding by paywalls, password protections, and the like,\u201d its lawyers wrote in the filing arguing Suno\u2019s training data was fair use. While Udio admits it may have used some copyrighted recordings, its response to the suit doesn\u2019t mention how they were acquired; if Udio bought those songs, under the Anthropic case\u2019s reasoning, it might be off the hook.<\/p>\n But that\u2019s not the only pertinent ruling. The very next day, in a case where authors alleged Meta had infringed on their copyright by training on their books, Judge Vince Chhabria directly addressed Alsup\u2019s ruling<\/a>, saying it was based on an \u201cinept analogy\u201d and brushed aside \u201cconcerns about the harm it can inflict on the market for the works it gets trained on.\u201d While Chhabria found in favor of Meta, he noted that it was because of bad lawyering on the part of the authors\u2019 team.<\/p>\n Still, the finding is better for music companies on the input side, because it doesn\u2019t draw a distinction around piracy, Grimmelman says. It is much, much worse for Suno and Udio on the output side. \u201cChhabria holds that \u2018market dilution\u2019 \u2014 creating lots of works that compete with the plaintiffs\u2019 works \u2014 is a plausible theory of market harm,\u201d he says in an email after the ruling. That\u2019s also in line with the copyright office\u2019s memo.<\/p>\n \u201cWe live in a world where everything is licensed.\u201d<\/p>\n Suno and Udio have some other trouble; some generative AI companies have been licensing artists\u2019 works. By offering nothing for works that other companies have licensed, they are messing up the market. \u201cThe fact that there are existing licensing deals for music training is relevant, if that market is better-developed than the market for licensing books,\u201d Grimmelman says. Chhabria\u2019s opinion points out that it\u2019s quite difficult to license books for training, because the rights are so fragmented. \u201cEither finding that there is a market that copyright owners should be able to exploit, or finding that there isn\u2019t one, is circular, in that the court\u2019s holding tends to reinforce its findings about the market.\u201d<\/p>\n That effectively stacks the deck against Suno and Udio, and any other music companies that didn\u2019t license their AI training data. Music licenses for AI training cost between $1 and $4 per track. High-quality datasets can cost from $1 to $5 per minute for non-exclusive licenses, and from $5 to $20 per minute for exclusive licenses. Transcription and emotion labeling, among other factors, garner higher prices.<\/p>\n And unlike in other industries, music already has an IP copyright and collection system, notes Kuok, of the BandLab recording app. The app has its own AI tool called SongStarter, which lets people who are making music begin with an AI-generated track. Kuok favors licensing music for AI training, and making sure musicians get paid.<\/p>\n \u201cWe live in a world where everything is licensed,\u201d Kuok says. \u201cThe solution is an evolution of what existed before.\u201d How to collect, who collects, and how much gets collected strikes Kuok as being open questions, but licensing itself is not. \u201cWe work in an all-rights-reserved world where we believe copyright is an important institution.\u201d<\/p>\n \u201cEveryone knew it was required.\u201d<\/p>\n To address that, BandLab has options for its licensing program. Artists can say they are open to AI licensing, which means they\u2019ll be contacted if a company wants to license their work. If they agree, their work is then bundled with an assortment of other artists\u2019 approved works for the licensing deal, which BandLab negotiates on their behalf. Kuok says Bandlab is discussing training deals now, though he declined to give specifics about the financial components of those deals, or who he was in talks with, <\/p>\n Kuok did say there were some other things he considers in negotiations. \u201cIt\u2019s important what the use is for,\u201d he says. \u201cThat has to be specified. These are fixed-term contracts, fairly large deals, worth six figures over a multiyear period.\u201d He recommends maintaining as much control as possible over copyrighted work to avoid diluting the value of existing IP.<\/p>\n That may be why Suno and Udio<\/a> are reportedly in talks with the majors to license<\/a> music for training their models. Other AI companies do already. Ed Newton-Rex, formerly of Stability AI, told me all the music he\u2019d worked with at Stability was licensed; he even quit his position<\/a> as a vice president at Stability after the company decided training on copyrighted data was fair use. He\u2019d been working on the systems since 2010, and licensing had been the norm until fairly recently, he told me. <\/p>\n \u201cEveryone knew it was the law,\u201d he says. \u201cEveryone knew it was required.\u201d<\/p>\n But after ChatGPT came out, some music AI companies thought they might also just grab whatever existed and let the courts sort it out. \u201cI don\u2019t think it\u2019s fair use,\u201d he says. \u201cGiven that gen AI generally competes with what it\u2019s trained on, it\u2019s a bad thing to take creators\u2019 works and outcompete them.\u201d Newton-Rex has also demonstrated ways to get Suno in particular to output music that\u2019s strikingly similar to copyrighted work<\/a>. That, too, is a problem.<\/p>\n \u201cI don\u2019t think there\u2019s an outcome where this winds up being all fair use,\u201d says Grimmelman.<\/p>\n