The comedian Bill Stiteler didn’t follow a conventional path to pop music stardom. In fact, he doesn’t really see himself as a musician at all.
A couple of years ago, Stiteler began creating jokey “realistic walk-up songs” for Major League Baseball players using Suno, an artificial intelligence music generator. The clips helped boost his follower count on Instagram. But his first certifiable hit came when he applied the same irreverent tone to travel and asked Suno to compose a song about a vacation trip.
His A.I.-assisted ode to an island, The Puerto Rico Song, has nearly 20 million Spotify streams, millions of views on YouTube and a licensing deal that earns him royalties (he won’t disclose dollar amounts). In June, Puerto Rico’s tourism board flew him out to record a music video. “It’s a wild place to vacation,” his A.I. singer belts. “Slot machines in the bus station.”
Stiteler, 37, has produced dozens of ticklish A.I. ditties about cities, often obscure ones (Altoona, Pa; Morgantown, W.Va; Manchester, N.H.) He writes the lyrics for what he calls his “pseudomusic”; Suno does the rest. “I’m not an artist,” he said. “I’m a fartist.”
The process gives him “such impostor syndrome about this musician stuff,” he added. “Musicians work their entire life. But I make this freaking A.I. Puerto Rico song and it’s smoking, like doing really well.”
The music industry itself has similarly mixed feelings about A.I. tools like Suno, which have set off a debate — and several lawsuits — over how to monetize the growing tsunami of A.I.-created tunes.
“Some folks believe that A.I. music should be eradicated,” said Rob Jonas, chief executive of Luminate, a music industry analytics company. “Others believe that music is not about how it’s made but how it makes you feel. If A.I. creates the same emotional reaction that organic music does, then it’s music.”
Suno (Hindi for “listen”) was founded in Cambridge, Mass., in 2022. Since then, around a dozen A.I.-generated songs have appeared on the Billboard charts. Deezer, a European music streaming platform, said last month that A.I.-generated songs accounted for half of its total uploads. Spotify, in partnership with Universal, is allowing users to pay to remix certain songs using A.I.
Over 100 million people have made music with Suno, the company says, each producing an average of 42 songs per month. Two million users pay up to $30 a month for more advanced features, like splitting a generated track into isolated vocals and instruments. For a few weeks this year, Suno was the most popular download in the music category of Apple’s app store. (Apple Music announced this month that it would soon require A.I. music to be tagged as such.)
There are competitors: Udio, Riffusion, ElevenLabs. But Suno, which has raised roughly $1 billion from the likes of Bond Capital and Union Square Ventures, with a $5.4 billion valuation, is the clear leader of this emerging industry. Many artists find the rise of such platforms scary. But in the telling of Mikey Shulman, Suno’s chief executive, no one should be afraid.
“The purpose of Suno is to allow the whole world to feel creative fulfillment,” he said. “In the future, this will be present in a lot of different domains. We do music.” (Not all artists are afraid: “It’s a new tool for creativity,” the hip-hop titan Dr. Dre told The New York Times this month. “I’m embracing it.”)
The labels and artists who see Suno as mostly a threat portray A.I. music less as another creative tool than another iteration of Napster-esque thievery. A hacker who breached the start-up and shared findings with 404 Media accused Suno of illegally scraping songs from YouTube Music, Genius, Deezer and others. (The hacked data “involved outdated source code that is no longer in use,” Suno told the publication. In a statement to DealBook, a spokesperson said that Suno’s models were “trained on publicly available music files on the open internet.”)
Even before the hack, major labels made similar accusations against Suno. In 2024, the Recording Industry Association of America, on behalf of Sony Music Group, Universal Music Group and Warner Music Group, sued Suno, accusing the company of training its models on copyrighted recordings without permission. Warner settled over a year later; Universal and Sony are still litigating. A parallel R.I.A.A. suit against rival platform Udio has since been settled by Universal and Warner, leaving Sony alone in court.
“Unlicensed services like Suno and Udio that claim that it’s ‘fair’ to copy an artist’s life’s work, and exploit it for their own profit without consent or pay, set back the promise of genuinely innovative A.I. for us all,” said Mitch Glazier, the head of R.I.A.A., in a statement at the time.
In any case, if you can’t beat them, join them. In November, Warner announced a partnership with its former litigant. Warner artists, from Ed Sheeran to Charli XCX, can now individually opt in to let Suno use their names, images, likenesses, voices and compositions in A.I.-generated songs. Suno also acquired Songkick, the concert-discovery platform, from Warner as part of the deal. Most significantly, Suno is currently developing a “licensed model,” trained on Warner’s catalog.
“We view our participation as the ability to shape the future and protect our artists and songwriters while at the same time creating new opportunities,” said the Warner Music Group chief executive Robert Kyncl. “The alternative is to just sit and watch.”
Much remains unclear in this first-of-its-kind deal. How do opted-in artists get compensated? What happens if a large share of artists opts out? In August, Suno reached a similar licensing deal with BMG. None of the companies have disclosed financial terms.
Some music businesses are choosing not to engage at all. In January, the indie streaming platform Bandcamp decreed an outright A.I. ban. The platform has no formalized system of identifying offenders. Rather, the company’s editorial director, J. Edward Keyes, said, You know it when you see it. “It really is the difference between sugar and artificial sweetener,” he said.
But millions use Splenda and Sweet’N Low. Will listeners enjoy artificial music, too? Will they pay for the pleasure? Luminate concluded in a recent report that a majority of Americans they polled “are tolerant — feeling either comfortable or indifferent of A.I. use in music they consume.” Billions of dollars are betting that remains so.
At the same time, said Jonas, “Whilst there may be a lot of A.I. music being created, there’s not a lot of it being listened to.”
After Stiteler sold the rights to “The Puerto Rico Song,” its new owners, the Universal-owned Xploded Music, commissioned a fully human remake. That is the version that now plays on Spotify. Stiteler’s Instagram page still features the A.I. original. If you listen carefully, you can hear the difference. Stiteler is pleased with the remake. “When musicians cover it,” he said, “it becomes even more awesome.”
IN CASE YOU MISSED IT
Kevin Warsh affirmed the Fed’s commitment to taming inflation. In a speech at the Fed’s annual conference in Jackson, Wyo., the chairman of the central bank said it would have “work to do” if price pressures did not ease, but stopped short of saying whether higher interest rates would be required. Short-term Treasury yields, which are sensitive to interest rate expectations, moved higher after the speech.
Meta agreed to pay up to $17.1 billion over social media addiction claims. In a settlement that ends a trial in which states sought roughly $200 billion from Meta, the company is initially going to pay $12 billion over the claims of 47 states (as well as the District of Columbia and U.S. territories) that its products were harmful and addictive to children. It will also make changes to its platform, including limiting teenagers’ use of Instagram and Facebook to two hours a day. The company agreed to pay an additional $5 billion only if its competitors Snap, TikTok and YouTube paid similar penalties and made similar changes.
This month, Representative Greg Casar of Texas, who chairs the Congressional Progressive Caucus, proposed a bill that would create such a tax. Sarah Kessler spoke with him about the plan and how A.I. has become a hot political topic.
At a basic level, how would it work?
We have two measurements of how much A.I. companies are selling. One is the tokens, which is the unit of measure they currently use to sell A.I. The other is revenue on A.I. sales. And to prevent them from gaming the system by rapidly changing the definition of a token, we tax the higher of those two.
The tax starts out quite low, when we don’t have A.I. mass unemployment, and then ramps up if A.I. starts to cause significant job loss.
People in the business community will say taxing A.I. tokens would slow down A.I use and make it less efficient. You’ve written that this would be a feature, not a bug. Why?
During the Industrial Revolution, we needed new child labor laws. In the nuclear revolution, we needed nonproliferation agreements and the Nuclear Regulatory Commission.
A.I. could be extremely disruptive to our economy. So the best thing we can do is get rid of the incentives to automate our jobs, and give workers the time we need to make sure that we don’t have mass unemployment.
But don’t we have to keep up with China?
Becoming more authoritarian like China, accepting major national security risks or being a country with mass unemployment to me does not sound like leading the world.
Many economists have argued that taxing an input to businesses, their A.I. tokens, is less efficient than raising corporate taxes, which would still allow companies to maximize their profits with A.I. before taxing them. Why is a token tax the best way?
By taxing A.I. directly, we tax something that would grow as unemployment grows, so we’re tying the solution to the problem.
If companies are using more A.I., we’re able to fund more and more jobs. Corporate profits could be related or unrelated to the A.I. itself.
Why funnel the tax income to support a new Work Protection Administration, which would fund new jobs, rather than broadly distribute the funds, which is favored by proposals from Senator Bernie Sanders and others?
I find a small universal basic income for everyone dystopic. Americans want to contribute. We want Americans to still be at work and still have a job.
I’ve been discouraged by the sole focus on job training as the response to mass unemployment. That would be like offering swimming lessons on the Titanic.
There are no federal A.I. regulations. Why?
Many of these A.I. billionaires, like Greg Brockman over at OpenAI, are threatening to spend unprecedented amounts of money in the midterm elections. And they’re trying to capture Trump’s Republican Party and to make the Democratic Party afraid to take them on.
A.I. is increasingly becoming a big issue in the midterm elections, especially as it relates to data centers. But I think it will move from a big issue to a very top-level issue in a 2028 presidential campaign.
Sports teams explode in value
When Jerry Buss bought the Los Angeles Lakers in 1979, The Times called it “the largest single financial transaction in the history of professional sports.” That deal was for about $68 million. Last year, the Buss family sold a controlling stake in the Lakers to financier Mark Walter for $10 billion, a record-shattering sum for a sports franchise. It was eclipsed earlier this month when Walter agreed to sell his share of the N.B.A. team to a group led by Bob Iger and Josh Kushner in a deal valuing the Lakers at $12.5 billion.
The Lakers are just one example of sports teams’ skyrocketing values. And recent deals underscore the big gains made over the past decade, Christine Zhang reports in the above chart. In July, the venture capitalist Vinod Khosla agreed to purchase the Super Bowl champion Seattle Seahawks for $9.6 billion from the estate of the Microsoft co-founder Paul Allen; Allen bought the N.F.L. team in 1997 for $194 million. Smaller franchises have seen their valuations soar, too: The N.B.A.’s Portland Trailblazers, also previously owned by Allen, were sold in March for $4.25 billion.
Whether or not they’re sold, team valuations are always changing. And usually rising. Real Madrid, the world’s most valuable soccer club, were estimated to be worth $9.5 billion by Forbes in May. That was a stunning 41 percent increase over the previous year, and reflected the club’s record $1.27 billion in revenue for its most recent fiscal year.
Thanks for reading! We’re taking a break next week and will see you again on Sept. 8.
We’d like your feedback. Please email thoughts and suggestions to dealbook@nytimes.com.
Follow DealBook on Instagram: @nytdealbook
