There’s a vibe shift going on in how the music industry is coping with its ongoing AI revolution, at least if you believe the man who is arguably the most responsible for starting that revolution: Mikey Shulman, CEO and co-founder of Suno, the music industry’s most prominent AI music generation platform.
“It wasn’t even happening at the end of last year, but in the past couple months since the beginning of this year,” Shulman says over zoom. “I don’t meet a lot of producers and songwriters who aren’t using Suno at least a little bit in their workflows. I think people are starting to be a little more comfortable being public and upfront about their use, and most importantly, I think a bit more optimistic about the future. It’s not everyone, but there’s definitely a market shift.”
Suno entered the industry as a major taboo, a musician killer trained without permission on millions of songs from the greatest recording artists of all time, an app that only needed a prompt to make a complete song. For creatives, its use was treated like blasphemy. By 2024, both Suno and its competing music generation platform Udio were sued by the Big Three major music companies in tandem, with the labels alleging mass copyright infringement.
And to be sure, that’s still a big part of the company’s identity. Two of the three major record labels remain in litigation with Suno — and many other industry stakeholders remain deeply skeptical. Just last month, a coalition of prominent artist advocates published an open letter called “Say No To Suno,” comparing the company to the thieves who made off with jewels at the Louvre last year. They deemed the model “the hijacking of the world’s entire treasure-trove of music.”
Still, there’s a sense the war could be deescalating. Udio carved out settlements and partnerships with Universal Music Group and Warner Music Group last fall, and Suno managed to settle with WMG back in November. (Sony remains in litigation with both companies, the only major label that hasn’t settled with either). With some of the labels now calling the AI labs partners, the platforms have looked to posture themselves not as invaders but as music companies themselves.
Indeed, while it remains a controversial topic, several songwriters who spoke with THR acknowledge that they’re seeing their peers use AI platforms like Suno more frequently.
Autumn Rowe, a career songwriter with credits on songs for Jon Batiste, Dua Lipa, Ava Max and many others, says many of her peers have used Suno to make demo productions of songs they wrote and have managed to get those songs placed with recording artists. (Once an artist takes it, the demo is produced by an actual producer, she says).
Rowe is an AI music skeptic herself but has started to lightly experiment with Suno. In recent weeks, she’s taken years-old demos she wrote that never got recorded and told Suno to remix them to see if they can be updated to try and give them a second life.
“I’ve got concerns with AI, I worry about younger writers who use Suno before they’ve spent the many, many hours crafting songs,” Rowe says. “But I do think AI in music will keep getting more prominent, and I think it could help writers get more leverage if they can do a lot of that production early themselves.”
***
As the music business continues to feel out the path ahead with AI, it provides a look into the not-so-distant future for the rest of the entertainment industry, which has largely moved at a slower pace.
Thus far in film and TV, the only studio with an active AI content deal is Lionsgate, which set up a deal with AI firm Runway to train a generative model on the studio’s IP. Disney announced a landmark deal with OpenAI’s Sora image generator at the end of 2025 that included a $1 billion investment from the studio, but exited the deal after OpenAI’s surprising announcement that it was shutting Sora down.
The major record labels have been much more active. Along with UMG and WMG’s respective deals with Suno and Udio, over the past few years the majors have announced myriad pacts with the likes of Spotify, Nvidia, Splice and Stability AI. On the other end, labels sued Suno and Udio — as well as AI giant Anthropic after finding the company used lyrics from its songs to train AI agent Claude — proving that they’ll take action when they feel the innovations cross the line.
Outside of the labels, Moises AI, a non-generative AI music service that provides tools for musicians (like vocal isolation and mastering) appointed Charlie Puth as its chief music officer earlier this month, with the singer serving as a consultant for the company’s creative and product direction.
It’s a marked change from the industry’s long-held reputation as a business of luddites resistant to change. Scars still exist from the Napster days and how unequipped the record labels were to handle the massive scale piracy from peer-to-peer filesharing sites, which cratered music’s value until the onset of the streaming era finally stopped the bleeding.
Tatiana Cirisano, a music industry analyst at Midia Research, says she’s “pleasantly surprised” with how the industry has handled AI thus far.
“The industry does have this reputation of being sort of resistant to tech disruption and not wanting to engage,” she says. “I’d say at least some of the response we’re seeing is the industry trying to prove that wrong.”
Ciricano says the music industry may actually be more equipped than film and TV to handle the AI era, at least for now, noting that AI is merely accelerating issues the music industry has already faced for years.
“The barrier for entry with music has been lowering and lowering for a very long time in the digital era, and that’s brought with it a lot of challenges and questions for the music industry already,” Ciricano says.
With the massive improvement of at-home recording hardware coupled with streaming becoming the primary means of consuming music, anyone can put up technically professional sounding music and put it on the same proverbial shelf as the world’s biggest artists. As many as 100,000 songs were getting uploaded to Spotify every day, even before services like Suno. YouTube has evened out the playing field for other content creators and stolen many eyeballs from traditional film and TV, though AI could represent another major shift as production on more ambitious film projects could get much cheaper for smaller creators outside of the film studios.
“AI speeds up the music challenges considerably, but some of the fundamental questions are the same,” she says. “That’s not something film and TV have faced the same way to this point. Music may be moving on from this a bit more.”
***
Since the WMG settlement, Suno seems to have gone into olive branch mode, even as it still faces suits from the two largest major record labels. During Grammy Week, Suno held a days-long songwriting camp at a recording studio in Hollywood, hosting industry executives, artists and songwriters with the goal of showing them how Suno could assist creators in their process.
“We were there and are there to listen a lot and learn and also show,” Shulman says of the camp.
The sessions could be startling for the uninitiated. The writers, led by grammy-winning producer Om’Mas Keith, fed lyrics to Suno in a prompt asking for different types of songs and vibes and, within minutes, had tracks with intricate production and convincing lead vocals. After Suno spit out the concept, a collection of world-class musicians including string players and a drummer recorded the musical bits to fill in the blanks and add a more personalized human touch.
For seasoned songwriters, Shulman pitches Suno as a supercharger.
“You hang around for one hour, it’s very apparent that the best creatives in the world make better shit with these things than us mere mortals do,” Shulman says.
And for those mere mortals, Shulman likens Suno to the gaming industry. He sees a future where creation is consumption. That could mean creating new music from scratch, or having more active ways to interact with other artist’s music. That latter concept has already been proven on platforms like TikTok, where users routinely speed up, slow down and remix songs for videos. Shulman teases features like that on the way for Suno but declines to divulge more.
“The whole world right now is passive consumption,” Shulman says. “But everybody is creative. Everybody has this drive to make something. In the future, people will be creating a lot more, and that means interacting with music in new ways. And of course, that means interacting with the music of their favorite artists in new ways.”
***
Rowe was among the songwriters who witnessed the Suno session during Grammy week. When asked for her thoughts on the camp, she said it was “interesting and there were some incredible people in the room, but I don’t know why people think we need [this tech].”
“I don’t know where the idea came from that this needs to move faster,” Rowe says. “The CEO of Suno can say people don’t like learning instruments or people don’t like the process of making music, but why is the music being made if it doesn’t come from a place of understanding or liking the process?”
Rowe is referencing what remains one of Shulman’s most controversial quotes, where he said in a podcast interview that he doesn’t think “the majority of people enjoy the majority of time it takes to make music.”
Shulman walked back that quote in an interview with Billboard earlier this month, saying that “I really wish I had chosen different words.” When I asked him to elaborate on what those words would be, he said that his intention wasn’t to imply that people don’t like making music and encouraged people to look back at his full quote.
“Yes, the struggle of making perfect music is actually enjoyable,” Shulman says. “But that doesn’t mean that we can’t improve the workflows and try to remove the tedium so that you can spend your creative minutes elsewhere. And I’m sure every single professional that you have talked to has bits of their workflow that they don’t find so enjoyable.”
Another major sticking point among AI critics is what value the generation tools are currently bringing into the ecosystem. The french streaming service Deezer reported earlier this year that it’s seeing 60,000 AI tracks getting uploaded to its platform every day, adding that as much as 85 percent of the streams on those tracks are fraudulent. Those findings would suggest an overwhelming majority of AI music consumption is powered by bad actors looking to syphon off royalties from legitimately streamed artists.
“We don’t have to theorize about the future of AI saturation as it’s become a marketplace reality. Most of this content is AI slop or fraud fodder,” UMG chief digital officer Michael Nash told analysts during the company’s recent earnings call in March, adding that AI brings promising opportunities for actual creation and fan interaction, but that fan interest in AI music on its own is minimal at best. “The aggregate organic consumption of AI content by actual consumers is less than half of 1 percent.”
Streaming services are acknowledging the increased fraud potential; as The Hollywood Reporter reported in January, Apple Music implemented a new policy to double penalties for those caught engaging in streaming fraud, with the platform’s head Oliver Schusser confirming the proliferation of AI content played a factor in the move.
Shulman pushes back on the fraud sentiment, calling the critiques a “sleight of hand” and lacking nuance. Suno, Shulman announced in February, had recently surpassed 2 million paying subscribers, which he argues reflects an interested user base, even if the songs aren’t hits.
“Even an AI track that the mass public doesn’t want to listen to is not valueless,” Shulman says. “I have so many Suno songs that I’ve made for myself with my kids that are not for mass consumption that I enjoy the hell out of, and that are extremely valuable to me. I think people need to be a little more hesitant before they pass judgment on those things.”
To say music has figured out the AI question would be a massive overstatement. As part of its settlement with Warner Music Group, Suno will be rolling out a new version of its model trained only on licensed music from that label at some point this year, and the old model will be phased out. Whether that new model will be nearly as effective as the one trained on millions of more songs remains to be seen. And it remains unclear how the artists and songwriters who agree to participate will specifically get paid.
Still, when pressed on those logistical issues, Shulman says he’s still optimistic. As he says: “You don’t need everybody at the beginning.”
