For more than a year since fake Drake track bam swarmed the music industry, industry executives have been debating whether generative artificial intelligence is a threat or an opportunity. The answer is both. Creators are already using AI tools to compete with AI music. But the future may be determined by a lawsuit filed two weeks ago by major record labels against Suno and Udio for copyright infringement, after the two companies allegedly used record labels’ music to train their AI programs.
Like most debates about technology, this one will be resolved in real time. Internet startups tend to find it easier to ask for forgiveness than to get permission. Neither Suno nor Udio say they trained their programs on music from major labels, but the labels’ lawsuits point out that both companies argue that using copyrighted works in this way is “fair use” and a defense to other infringement claims. The companies aren’t admitting that they did it, just defending themselves in case they did.
Whether this constitutes fair use is a million-dollar question, as statutory damages could reach $150,000 per infringed work. But the risks go beyond that: if fair use allows for the ingestion of copyrighted material en masse to train an AI, the music industry could find it difficult to restrict, control, or monetize the technology.
If that doesn’t happen, the record companies will have at least some control over these companies, and possibly the entire emerging sector. There are other ways to restrict AI, from legislation to publicity rights, but copyright law is the only one with statutory damages that will have any real impact.
Neither Snow nor Woody have filed any legal action, but Snow’s CEO Mikey Schulman The company released a statement saying that record companies have “returned to their old lawyer-driven tactics” – the obvious reference being Napster, since most people think the music industry saw its future in the late ’90s and decided to sue Napster.
It wasn’t. The big record companies knew the future was digital. They lobbied for the passage of the Digital Recording Performance Rights Act of 1995, which required streaming services to pay a fee so that traditional radio stations in the US didn’t have to pay to play recordings. They just didn’t want peer-to-peer services to distribute their content for free, and they had to negotiate while they were distributing it. In July 2000, three months after the big record companies sued Napster, key executives met with Napster and couldn’t come to an agreement. The record companies negotiated as if Napster needed a license, and Napster negotiated as if it didn’t. Eventually, after a decade of litigation and lost business, creators and rights holders established their right to be paid for online distribution, and the music industry began to recover.
And then it happens again. History doesn’t repeat itself, but it does seem to rhyme. If the labels negotiated with Suno or Udio now, how much would those companies be willing to pay for rights they may or may not need? It’s easy to kid either side, but it’s hard to know how much to charge or pay for rights before you even know if you need them.
These lawsuits are not about whether creators and rights holders embrace or shun AI. AI is coming for better or for worse. In modern terms, the question is whether embrace is consensual, and on what terms. Most creators and rights holders want to do business with AI companies, so long as that means practical things. work — negotiate a deal in something resembling a free market.
Their fear is that tech companies will build empires using their work without paying them, and especially that they will create products that produce music that competes with theirs. That depends on the outcome of these lawsuits, because without the right to say no, you can’t get a fair yes.
A few weeks ago, at a cultural conference organized by the German record music industry association, the German Minister of Justice Marco Buschmann Buschmann put it better than anyone I’ve ever heard him say it: “The moment people have the opportunity to say ‘no,’ and the moment they can enforce that ‘no,’ they have legal negotiating position,” he said in his speech. (Coincidentally, Buschmann also makes electronic music.) In the European Union, rights holders can veto AI incorporation, which is far from ideal, but better than nothing.
What happens in the United States, which often shapes the global media business, may depend on the outcome of these cases. There are 24 copyright cases regarding AI, but these seem to be among the most important. Some of the others are caught up in jurisdictional games, while others are less powerful. The New York Times Fair use determinations can change if the imported article is used as a source and not used to create a new work. These cases are simple, but they are not going to move quickly. It is easy to see this issue making it to the Supreme Court.
Despite the high stakes, and the near certainty of a back-and-forth that will rival the rap debate, fair use judgments are quite nuanced. Fair use can be legal to use excerpts or the entirety of a copyrighted work without permission, usually for commentary purposes. (An iconic Supreme Court case involved 2 Live Crew’s parody of Roy Orbison’s song “Oh, Pretty Woman.”) This is a long way from that, but Suno and Udio will likely argue that their actions qualify as “transformative use,” just like the Google Books project. Next week, I’ll write about the arguments we can expect to hear, the rulings we might see, and what could happen while we wait for them.
