artificial intelligence
The three major music companies — Sony Music, Universal Music Group and Warner Music Group — are in talks with AI music companies Suno and Udio to license their works as training data, despite suing the two startups for infringement “on an almost unimaginable scale” last summer. Now, executives in the “ethical” or “responsible” AI music space are voicing displeasure that the alleged infringers could potentially benefit from their actions.
Several of those ethical AI companies said they were led to believe they would be rewarded by the record labels for going through the tough process of licensing music from the beginning, in what one AI music company founder previously told Billboard would be “a carrot and stick approach to AI,” penalizing those who raced ahead and trained models without permission or compensation.
Trending on Billboard
“That’s all out the window,” that founder says now. “I was talking to another founder that does ethical AI voice models, and he told me, ‘F–k it. I don’t care anymore. Why does it matter to be ethical if we just get left behind?’”
Ed Newton-Rex, founder of non-profit Fairly Trained, which certifies ethically-trained AI models, adds: “If I were running a startup that had tried to do the right thing — respecting creators’ rights — and investors had rejected me because I wasn’t exploiting copyrighted work like so many others, and then this happened? I’d definitely be pissed off.”
Tracy Chan, CEO of AI music company Splash, told Billboard via email that she stands by her decision to license music from the start. “At Splash, being ethically trained wasn’t a debate — it was obvious,” she says. “We’re musicians and technologists. We believe AI should amplify creativity, not exploit it. We don’t need to scrape the world’s music to make that happen.”
It remains unclear how far along these licensing talks are between the major music companies and Suno and Udio, and if deals will even come to fruition to avert the blockbuster lawsuits. It’s common in costly and lengthy litigation like this for the two sides to discuss what it would look like to settle the dispute outside of court. Plus, licensing is what the majors have wanted from AI companies all along — does it matter how they come to it?
Multiple executives expressed fear that if the majors ditch the lawsuit and go for deals, they will set a bad precedent for the entire business. “Basically, if they do this deal, I think it would send a message to big tech that if you want to disrupt the music industry, you can do whatever you want and then ask for forgiveness later,” says Anthony Demekhin, CEO/co-founder of Tuney.
This, however, is not the first time the music business has considered a partnership with tech companies that were once their enemy. YouTube, for example, initially launched without properly licensing all of the music on its platform first. In his 2024 New Years’ address to staff, Lucian Grainge, CEO/chairman of UMG, alluded to this, and how he would do it differently this time with his so-called “responsible AI” initiative. “In the past, new and often disruptive technology was simply released into the world, leaving the music community to develop the model by which artists would be fairly compensated and their rights protected,” he wrote, adding that “in a sharp break with the past,” UMG had formed a partnership with YouTube to “give artists a seat at the table” to shape the company’s AI products, and that the company would also collaborate “with several [other] platforms on numerous opportunities and approaches” in the AI space.
Another part of Grainge’s “responsible AI” initiative was “to lobby for ‘guardrails,’ that is public policies setting basic rules for AI.” Mike Pelczynski, co-founder of ethical AI voice company Voice-Swap, also worries that if these deals go through, they could weaken the music industry’s messaging to Capitol Hill, where bills like the NO FAKES Act are still in flux. “All the messaging we had before, all the hard-lining about responsible AI from the beginning, it’s gone,” he says. “Now, if policy makers look at [the music business] they might say, ‘Wait, what side should we take? Where do you stand?’”
If talks about licenses for Suno and Udio move forward, determining exactly how that license works, and how artists will be paid, will be complex. To date, almost all “ethical” AI companies are licensing their musical training data from production libraries, which offer simple, one-stop licenses for songs. Alex Bestall, CEO of music production house and AI company Rightsify, says that the structure of those deals are typically “flat-fee blanket licenses for a fixed term, often one to three years or in some cases perpetuity… all data licensing [music or otherwise] is pretty standardized at this point.”
It’s unclear if the deals the majors have discussed with Suno and Udio will follow this framework, but if they did, the question then comes — how do the majors divide up those fees for their artists and writers? The Wall Street Journal reported that “the [music] companies want the startups to develop fingerprinting and attribution technology — similar to YouTube’s content ID — to track when and how a song is used.” In that scenario, the money received would be given to signees based on usage.
While there are a few startups working on music attribution technology right now, multiple experts tell Billboard they don’t think the tech is ready yet. “Attribution is nowhere,” says Newton-Rex, who also previously worked as vp of audio at Stability AI. “It’s not even close. There’s no system that I have seen that would do a decent job of accurately assigning attribution to what has inspired a given song.”
Even the possibility of deals between the parties has sparked a larger conversation about how to handle tech companies who ask for forgiveness — and not for permission — from the music business.
“If the two biggest offenders actually become the legal standard, it’s effectively like making Pirate Bay into Spotify,” says Demekhin. “I understand it from a business perspective because it’s the path of least resistance [to settle and get a license now]. But this could send a message to tech that could bite the industry on the next wave.”
Sir Elton John collected the Creators’ Champion Award at Billboard’s Global Power Players event in London on Wednesday (June 4) and used his speech to warn the U.K. government that “we will not back down” in relation to its controversial AI data bill.
Joined by his manager and husband David Furnish, John was awarded the inaugural trophy by Billboard editor-in-chief Hannah Karp in recognition for standing up for the creative industry and calling for fair usage and protections for music rights holders in relation to AI models.
As referenced in John’s speech, the Labour government’s data bill – which proposes an opt-out approach for rights holders, as opposed to the preferred opt-in strategy – has been voted down by the House of Lords for an unprecedented fifth time. Led by Baroness Kidron, the ping-pong between the Houses of Commons and the upper house has lasted for weeks and resulted in John calling the government “absolute losers” in a recent interview.
Trending on Billboard
Taking to the stage, John thanked Billboard for the accolade as he was honored alongside other industry leaders such as Empire’s Ghazi and Tems on the night. “Supporting the next generation of British artists is the one of the major driving forces in my life,” he said and thanked the Lords who backed the amendment saying “you rock.”
He continued: “The onus is now on the government to do the right thing and have transparency added to the bill. We’ve asked them so many times to come to us. We’ve had so many meetings with them saying that we wanted to work with you.
“Copyright has to be transparent and seek permission,” he continued. “These two principles are the bedrock of the industry and they must be included in the data bill as a backstop. Let’s be clear: We want to work with the government, we want our government to work with us. We are not anti-AI and we are not anti-Big Tech.”
John added in his passionate speech: “We are not against Labour and we want a solution. We want to bring all parties together in a way that is transparent and fair and allows artists to maintain control of their work.” He concluded: “We will not back down and we will not go away quietly. This is just the beginning.”
Read the full speech below.
Elton John’s full speech at Billboard‘s Global Power Players event
“Thank you to Billboard for honoring myself and David with this incredible award. Supporting the next generation of British artists is one of the major driving forces in my life. It’s so important. As everyone in this room is aware, the data bill is currently looming over us and our musicians and our industries and the future of the livelihood of all artists.
It is an existential issue. In the last few hours the government has been defeated for an unprecedented fifth time by the House of Lords – that’s never happened before. The Lords, who have backed our crucial amendment to the bill – thank you so much – you rock.
The onus is now on the government to do the right thing and have transparency added to the bill. We’ve asked them so many times to come to us. We’ve had so many meetings with them saying that we wanted to work with you. I voted for you, you promised me that you wanted to get young artists into Europe. I did a show for you and to get more finance into the country and investment. I did a show at St Paul’s Cathedral… and we’ve heard nothing.
Copyright has to be transparent and seek permission. These two principles are the bedrock of the industry and they must be included in the data bill as a backstop. Let’s be clear: we want to work with the government, we want our government to work with us. We are not anti-AI and we are not anti-Big Tech.
We are not against Labour and we want a solution. We want to bring all parties together in a way that is transparent and fair and allows artists to maintain control of their work. It’s so fucking important. A machine is not a human being; it doesn’t have a soul. An artist when they create or they write or they sculpt or they paint or they photograph has a soul, it’s coming from the feeling of a person. And it’s so important if we ever lose that, we are well and truly fucked.
We will not back down and we will not go away quietly. This is just the beginning and thank you Billboard and Baroness Kidron for standing up for our world beating artists and journalists and playwrights and authors.
Let’s put it like this: The entertainment industry brings £125 billion to the United Kingdom and is the second biggest industry in this Great Britain. And they are treating us like shit. They are treating us like idiots. They have their head in the sand. The Labour government at the moment should be called the Ostrich Party. I will fight for this until it’s done and people have a fair deal. Every young artist or new person who writes a song, I want them to have a future and have their copyright respected.
There’s 2.4 million people in this industry in Britain, and 70% of them live outside of London. This isn’t a case of wealthy London people, this is a nationwide thing. Please listen to us, government. If you don’t, there’s going to be trouble. If you don’t, I will come banging on your door. But let’s get together: give me a call! Let’s have a cup of tea! So far, you haven’t responded to anything we’ve suggested. Thank you!”
ABBA‘s Björn Ulvaeus is working on a new musical using artificial intelligence. According to Variety, during a talk at SXSW London on Wednesday (June 4), the 80-year old Swedish pop legend said he’s tapping into AI because he believes it is an excellent creative tool.
Explore
See latest videos, charts and news
See latest videos, charts and news
“Right now I’m writing a musical, assisted by AI,” Ulvaeus said, noting that he’s about three-quarters of the way through the creative process on the unnamed project, the follow-up to the hugely successful pop quartet’s avatar stage show, Voyage.
“It’s fantastic. It is such a great tool,” Ulvaeus raved of AI. “It is like having another songwriter in the room with a huge reference frame. It is really an extension of your mind. You have access to things that you didn’t think of before.” Unlike many in the industry who fear that AI is an existential threat to their existence and the traditional creative process, Ulvaeus is aware of the bugs in the system, which he said have helped him to merge AI with his already formidable songwriting skills.
Trending on Billboard
“It’s lousy at [writing a whole song]” and “very bad at lyrics,” he said about his AI helpmate, which has allowed him to navigate through some creative dead-ends. “You can prompt a lyric you have written about something, and you’re stuck maybe, and you want this song to be in a certain style,” he said. “You can ask it, how would you extend? Where would you go from here? It usually comes out with garbage, but sometimes there is something in it that gives you another idea.”
Ulvaeus is part of an eclectic lineup for 2025 SXSW London, whose lineup includes Erykah Badu (as DJ Lo Down Loretta Brown), Tems, Mabel, Alice Glass and many more. Penske Media Corporation (which also owns Billboard) and the film and production company MRC invested in SXSW in 2021 following the challenges of the COVID-19 pandemic; two years later, Penske took majority ownership of SXSW.
The AI project, whose ultimate form has not yet been announced, is part of Ulvaeus’ ongoing partnership with Pophouse Entertainment, the company behind the ABBA Voyage production. The Voyage virtual residency opened in London in May 2022 and is slated to run through January 2026. The show is a combination of 10 live performers and digital avatars of the four ABBA members, who have not performed live since their split in December 1982; the group, which also features Agnetha Fältskog, Benny Andersson and Anni-Frid Lyngstad, released their first album in 40 years, Voyage, in November 2021.
Though Ulvaeus is happy to use AI in the creative process, he is also adamant about fighting for artists’ rights in the rapidly evolving digital age. “These AI models wouldn’t exist without the songs that we wrote,” he said.
THE BIG STORY: Is the music industry’s billion-dollar legal battle over artificial intelligence suddenly going to be over before it starts? I wouldn’t count on it just yet.
News broke this weekend that Universal Music, Warner Music and Sony Music were each in talks to license their songs to Suno and Udio for use in training AI models to spit out new songs – less than a year after the music companies filed blockbuster copyright cases accusing the AI startups of stealing music on an “unimaginable scale.”
Both sides have framed those cases as an existential fight over the future of music. The labels says Suno and Udio are “trampling” the rights of real musicians in an effort to replace them; the startups argue back that the music giants are abusing intellectual property to crush a new technology that threatens their market share.
Trending on Billboard
Is an actual deal imminent? I’d lean toward no. Low-boil settlement talks are a common and continuous feature of almost any litigation, and it’s unclear exactly how far they’ve gotten here. The labels have long said they’re open to listening, but striking a final deal – rather than just discussing one in broad strokes – will require solving a long list of incredibly complex problems, as my colleague Kristin Robinson writes.
The choice confronting the music giants in these talks – to strike a deal or fight it out in court – is one facing all creative industries amid the rise of artificial intelligence. If you sell your content to AI firms, you earn short term profits…by empowering a technology that might destroy you in the long run. If you fight it out, you risk being left behind amid a technological revolution, all for what could be a losing cause.
For the music business, history makes that an especially grueling choice. Nobody wants a replay of the 2000s, when labels chose to fight an ascendant new technology in court rather than harness and exploit its potential for profit. Only when the majors opted to partner with streamers like Spotify – in equity stake deals eerily similar to those being discussed with Suno and Udio – did the industry begin the long climb to recovery.
But digital music merely represented a new distribution system, and one where the tech partners would ultimately always need to pay real musicians for their output. If the labels end these lawsuits by handing over their catalogs for AI training now, those new partners might one day no longer need them at all.
You’re reading The Legal Beat, a weekly newsletter about music law from Billboard Pro, offering you a one-stop cheat sheet of big new cases, important rulings and all the fun stuff in between. To get the newsletter in your inbox every Tuesday, go subscribe here.
Other top stories this week…
LIZZO’S APPEAL – Lizzo’s lawyers launched an appeal aimed at ending a sexual harassment lawsuit filed by her former backup dancers, calling it an “attack” on her “First Amendment right to perform her music and advocate for body positivity.” Lizzo’s attorneys say her behavior toward the dancers was clearly part of her artistic approach — and thus shielded by constitutional protections for free speech.
SHADY FACEBOOK MUSIC? – Eight Mile Style, a company that owns much of Eminem’s catalog, filed a copyright lawsuit against Meta over accusations that Facebook and Instagram made “Lose Yourself” and other iconic tracks available to billions of users without permission. The case claimed that the social media giant added Eminem’s songs to its music library without the necessary licenses, violating copyright law on a “massive” scale in the pursuit of “obscene monetary benefit.”
OFFENSE AS DEFENSE – Smokey Robinson filed a countersuit against four longtime housekeepers who recently accused him of rape, claiming the allegations were part of an “extortionate scheme” by the women and their attorneys. The Motown legend accused the women and their lawyers of defamation, invasion of privacy, civil conspiracy and even elder abuse over the “fabricated” allegations.
YOUNGBOY PARDONED – President Donald Trump granted a pardon to YoungBoy Never Broke Again, who was released from prison in April after pleading guilty last year to a single count of possession of firearms by a convicted felon. The rapper, who has faced legal trouble for years, thanked the president on social media: “This moment means a lot. It opens the door to a future I’ve worked hard for and I am fully prepared to step into this.”
TRAFFIC DEATH – Ex-Red Hot Chili Peppers guitarist Josh Klinghoffer reached a plea deal with prosecutors to avoid prison time after striking and killing 47-year-old Israel Sanchez with his car in Los Angeles last year. Under the terms of the deal, the rocker pleaded no contest to misdemeanor vehicular manslaughter without gross negligence and was sentenced to one year of informal probation and 60 days of community labor.
SPOUSAL SUPPORT – Offset updated his divorce filings to demand that his estranged wife Cardi B pay him spousal support after their split is finalized. In an amended version of his answer to Cardi’s divorce petition, the Migos star added a request for an unspecified amount of alimony, but remains unchanged otherwise. The pair of superstars are one year into an increasingly acrimonious divorce case following six years of marriage.
BILLION WITH A B – The Justice Department urged the U.S. Supreme Court to tackle a billion-dollar lawsuit over music piracy filed by the major labels against Cox Communications, warning that a “sweeping” ruling could force internet providers to cut off service to many Americans. The case, in which the labels won a $1 billion verdict in 2019, saw a lower court hold Cox itself liable for widespread illegal downloading by its users.
FIVIO PLEA DEAL – Brooklyn rapper Fivio Foreign took a plea deal to end a criminal case stemming from allegations he pulled a gun on a New Jersey woman after she asked him to jump her car while pulled over. Under the terms of the deal, Fivio admitted to one count of third-degree terroristic threats in return for prosecutors dropping four other charges, including unlawful possession of a weapon and aggravated assault.
“A Million Colors” by Vinih Pray has become the first-known AI-generated song to hit the TikTok charts. Currently sitting at No. 44 on the TikTok Viral 50, the doo-wop inspired song was generated using the popular AI music platform Suno, the company has confirmed.
While this marks the first AI-generated song to hit the TikTok charts, there was one previous song on the TikTok charts that was human-made but contained an AI-generated sample. In the summer of 2024, “U My Everything” by Sexyy Redd, featuring Drake — which sampled the AI-generated song “BBL Drizzy” — peaked at No. 2 on the TikTok Top 50, but the most common sound clip used from “U My Everything” did not feature that sample.
With over 371,700 creates on TikTok, 819,745 streams on Spotify, 161,000 views on YouTube, “A Million Colors” sounds so realistic that it has fooled a number of unsuspecting users. One of the most popular celebrities on the platform, Kylie Jenner, recently posted a makeup tutorial with the AI-generated song in the background, earning her 1.5 million likes.
Trending on Billboard
Some TikTok users, however, have started catching on. On “A Million Colors” sound page, a number of videos are being made by users to call out its use of AI. This likely traces back to a popular video by @americangorls, who wrote in a post on May 10 “this song having hundreds of thousands of uses and I haven’t seen anyone talking ab the fact that this is 100% ai is freaking me out a little. am i crazy[?]”
According to his Spotify bio, Pray, the artist who posted “A Million Colors,” was born in Rio de Janeiro, Brazil and is “making waves in the Brazilian music scene,” and that “in addition to his solo career, [Pray] has collaborated with several renowned artists,” although it does not state who those artists are. He releases music at a quick pace — since he started posting to the page in April 2024, Pray has released over 110 songs to Spotify and other platforms in a variety of languages, including Korean, English, Portuguese and Mandarin. Most songs contain different voices, and it is unclear if any are using Pray’s own voice. It is also unclear how many of these songs are AI-generated or human-made.
Pray and TikTok have not replied to Billboard’s request for comment.
With “A Million Colors,” Pray positions himself among a growing class of AI music content creators on the internet. King Willonius, the artist who used Udio to generate “BBL Drizzy” posts a number of AI-generated parody songs to his 56,000 followers on Instagram. Another popular AI music poster, AI For the Culture, has garnered 150,000 subscribers on YouTube for his work, which includes AI-generated period music, paired with an AI image and a fictional artist bio to humanize the track. One of his songs, “Turn On The Lights” — which used AI to reimagine a Future’s “Turn On The Lights” as 1970s soul — was sampled in JPEGMAFIA‘s “either on or off the drugs” last year.
Over the weekend, Bloomberg broke the news that the Sony Music, Universal Music Group and Warner Music Group are in talks with Suno and Udio to license their music to the artificial intelligence startups. If the deals go through, they could help settle the major music companies’ massive copyright infringement lawsuit against Suno and Udio, filed last summer.
Billboard confirmed that the deals in discussion would include fees and possible equity stakes in Suno and Udio in exchange for licensing the music — which the two AI firms have already been using without a license since they launched over a year ago.
That sounds like a potentially peaceful resolution to this clash over the value of copyrighted music in the AI age. But between artist buy-in, questions over how payments would work and sensitivities on all sides, the deals could be harder to pull off than they seem. Here’s why.
Trending on Billboard
You need everyone on board
Ask anyone who’s tried to license music before: it’s a tedious process. This is especially true when a song has multiple songwriters, all signed to different companies — which is to say, almost all of pop music today. Since any music that is used as training data for an AI model will employ both its master recording copyright and its underlying musical work copyright, Suno and Udio cannot stop at just licensing the majors’ shares of the music. They will also need agreements from independent labels and publishers, too, to use a comprehensive catalog.
And what about the artists and songwriters signed to these companies? Generative AI music is still controversial today, and it is foreseeable that a large number of creatives will not take too kindly to their labels and publishers licensing their works for AI training without their permission. One can imagine that the music companies, to avoid a revolt from signees, would allow talent to either opt-out of or opt-in to this license — but as soon as they do that, they will be left with a patchwork catalog to license to Suno and Udio. Even if a song has one recording artist and five songwriters attached to it, it only takes one of those people to say no to this deal to eliminate the track from the training pool.
Is the expiration date really the expiration date?
Licensing music to train AI models typically takes the form of a blanket license, granted by music companies, that lasts between one and three years, according to Alex Bestall, CEO of Rightsify, a production music library and AI company. Other times it will be done in perpetuity. Ed Newton-Rex, former vp of audio for Stability AI and founder of non-profit Fairly Trained, previously warned Billboard that companies that license on a temporal basis should look out for what happens when a deal term ends: “There’s no current way to just untrain a model, but you can add clauses to control what happens after the license is over,” he said.
Attribution technology seems great — but is still very new
Many experts feel that the best way to remunerate music companies and their artists and songwriters is to base any payouts on how often their work is used in producing the outputs of the AI model. This is known as “attribution” — and while there are companies, like Sureel AI and Musical AI, out there that specialize in this area, it’s still incredibly new. Multiple music industry sources tell Billboard they are not sure the current attribution models are quite ready yet, meaning any payment model based on that system may not be viable, at least in the near term.
Flat-fee licenses are most common, but leave a lot to be desired
Today, Bestall says that flat-fee blanket licenses are the most common form of AI licensing. Given the complexities of fractional licensing (i.e., needing all writers to agree) with mainstream music, the AI music companies that are currently licensing their training data are typically going to production libraries, since those tend to own or control their music 100%. It’s hard to know if this model will hold up with fractional licensing at the mainstream music companies — and how they’ll choose to divide up these fees to their artists.
Plus, Mike Pelczynski, founder of music tech advisory firm Forms and Shapes and former head of strategy for SoundCloud, wrote in a blog post that “flat-fee deals offer upfront payments but limit long-term remuneration. As AI scales beyond the revenue potential of these agreements, rights holders risk being locked into subpar compensation. Unlike past models, such as Facebook’s multi-year deals, AI platforms will evolve in months, not years, leaving IP holders behind. Flat fees, no matter how high, can’t match the exponential growth potential of generative AI.”
There’s still bad blood
The major music companies will likely have a hard time burying the hatchet with Suno and Udio, given how publicly the two companies have challenged them. Today, Suno and Udio are using major label music without any licenses, and that defiance must sting. Suno has also spoken out against the majors, saying in a court filing that “what the major record labels really don’t want is competition. Where Suno sees musicians, teachers and everyday people using a new tool to create original music, the labels see a threat to their market share.”
Given that context, there is a real reputational risk here for the labels, who also represent many stakeholders with many different opinions on the topic — not all of them positive. For this licensing maneuver to work, the majors need to be able to feel (or at least position themselves to look like) they came out on top in any negotiation, particularly to their artists and songwriters, and show that the deals are in everyone’s best interests. It’s a lot to pull off.
Spotify has launched a new masterclass to help artists understand what artificial streaming is and how to prevent them from falling for scams related to it. Featuring Bryan Johnson, head of artist and industry partnerships, international at Spotify; Andreea Gleeson, CEO of TuneCore and member of the Music Fights Fraud Alliance; and David Martin, CEO of the Featured Artists Coalition, the executives also tell artists in the masterclass what to do if they’ve noticed abnormal and suspicious streaming activity has occurred on their accounts.
In recent years, artificial streaming (sometimes known as ‘streaming fraud’) has become a hot topic in the music business, but last year, the issue hit new heights when the first-ever U.S. streaming fraud case was brought against a man named Michael Smith in the Southern District of New York. According to the lawsuit, Smith allegedly stole more than $10 million in royalties across all streaming platforms by uploading AI-generated songs and driving up their stream counts with bots.
Trending on Billboard
A lot of artificial streaming instances, however, are not so extreme — or deliberate. It’s an issue that can impact even well-meaning independent artists, looking earnestly for marketing and promotion help. Whether it’s a digital marketer on Fiverr promising to get an artist on a playlist, or a savvy promoter DMing an artist on Instagram, promising a certain number of streams, many artists, particularly those without representation, fall for a scheme which, as Gleeson puts it, “is too good to be true.”
“It undermines the fair playing field that streaming represents,” says Johnson in the masterclass. “If left unchecked, artificial streams can dilute the royalty pool and shift money away from artists who are genuinely trying to release music and build an audience, and it can divert that money to bad actors looking to take advantage of system.”
As Johnson explains in the masterclass, whether the artificial streaming scheme was done wittingly or not, there can be real consequences. If an artist is caught with suspicious streaming activity, the track can no longer earn royalties, future streams do not count toward public metrics, future streams do not positively influence recommendation algorithms, and the activity is reported to the distributor or label. In a worst-case scenario, the song can also be removed from Spotify playlists, or the platform overall, and the artists’ label or distributor will be charged a fine.
“This is something we take seriously at every level, all around the world, and our efforts are working. Less than 1% of all streams on Spotify are determined to be artificial,” says Johnson.
Johnson, along with Gleeson, spoke to Billboard to explain why they teamed up on this masterclass, and what they hope artists take away from it.
“Our tactics are working, but this is not the time for us to pause. We have to keep going, because this is a moving target. We have to keep investing, keep educating, and keep trying to minimize the impact of artificial streaming,” says Johnson.
Why did you decide to do this masterclass now?
We have an existing video on artificial streaming from a few years ago, but we think this is a great way to update what we’ve done previously. This moves so quickly, and we will keep building on this and updating it. This masterclass now includes information that wasn’t around a few years ago. Why now? We found it’s very clear on our socials that people want to know about artificial streaming. We’ve read through a lot of comments, talked to industry partners and tried to figure out what information people want to see from this. And how can we do this in a way that will deter people from falling for this in the first place?
If artists only took one thing away from this video, what would you want them to have learned?
Gleeson: The big thing that that we really stress is that artists should know who they’re working with. If it sounds too good to be true, it probably is. So do your education up front with the marketing programs that you are utilizing. If a company is guaranteeing playlist placements or a certain number of streams or it sounds just way too good to be true, it probably is. If I can add a second thing it would be: start to open up a proactive discussion with your distributor or your label to make them aware if you see anything abnormal happen to your account. This can help you reduce any penalties, like the risk of your music getting taken down. Spotify also now has a way to submit information if you think there’s abnormal activity on your account. Reporting it is really important.
Bryan, often in these conversations about artificial streaming, it’s hard to know where the buck should stop. Is it the distributors fault? The user’s fault? The streaming service’s? How are you approaching it at Spotify?
Johnson: I think it’s an industry responsibility. There’s an industry body called Music Fights Fraud Alliance, which is a collection of digital services, like Spotify, and rights holders, like distributors and labels. And it’s an opportunity for the industry to come together and rally around the same topic and share information, share intel. And it’s been highly effective — super, super productive. We have a responsibility. We are the leading streaming service globally. We are across 184 markets with a huge audience, and we are a significant partner to the music industry. So it’s important that we come to the table, and we’re part of this conversation.
Scammers evolve quickly over time. It’s often said that there’s a danger in educating the public too much about artificial streaming, for fear that it will help the scammers evolve and better their efforts. How did you guys approach that challenge with this master class?
Gleeson: It’s always a work in progress to figure out the right balance of education and secrecy. You want to be specific to help educate, but you don’t want to give someone a playbook of how to do the fraud either.
From your vantage point, what are some of the measures that have been have implemented over the last few years that have shown really strong results in decreasing artificial streams? What has been the most effective?
Gleeson: We’ve tried to take down some of the playlists where this is happening. That can be a little bit of Whack-a-Mole, but that has been really effective. What we’ve observed also is the [streaming services] that focus on the reporting tools for artists and preventative tools have much lower abnormal stream levels than other [streaming services]. And it makes sense, right? If you’re a bad actor, you’re going to go to the path of least resistance. If a [streaming service] hasn’t invested yet in a robust reporting system, you can do your scam there and achieve more with less effort.
Johnson: We launched the very effective tool around a year ago, and it’s essentially a form that artists can go to, and they can tell us if they’ve seen abnormal activity, like in their Spotify For Artists data, and if they think they’ve been added to some sort of suspicious playlist. This born out of their feedback. We were getting requests online for a way in which they could let us know directly. Yes, they let their label and distributor know, but they also wanted to let Spotify know directly, and it’s super useful for us. We’re able to use that information effectively to stop artificial activity. So I think that playlist reporter form has been, has been really useful.
AI music company Suno has debuted a number of new features, allowing users to have more control over the creation and customization of their songs. The company has been known as a powerful generative tool that can make realistic songs at the click of the button with just a few simple prompts, but now, it is embracing more of a collaborative approach between the user and its AI technology.
Now, users can upload up to eight minutes of audio, whether its a hummed melody or a mostly-completed track, and then use Suno to remix it or expand it. Through their new Song Editor features, users can replace lyrics and reimagine sections of songs, as desired. Offering what it calls the “creative slider” users can rework a song into a new genre by using simple toggles that can up a song’s “weirdness” “style strength” or “audio strength.”
The company has also integrated a stem extraction tool to split a Suno-generated song into 12 clean stems, which can then be exported to a user’s preferred DAW.
Trending on Billboard
Suno CEO/founder Mikey Shulman says of the new features: “We envision Suno as a core part of musical creativity, for everyone from novices to Grammy winners. Our new tools offer powerful ways to explore new sonic ideas, remix, and iterate. Our upgraded editing suite gives artists more control than ever over their music.
We’re also working to better integrate Suno into the music production process. For example, Suno’s new stem extraction feature allows artists to more easily bring what they do with Suno into their favorite DAW, and vice versa. We’re excited to keep building tools like these to augment the creativity of musicians and expand access to serious music making.”
The announcement comes just a couple days after news broke that Suno, and its competitor Udio, are in talks with the major music companies — Sony Music, Universal Music Group and Warner Music Group — about licensing their copyrights for AI training. This may include the majors receiving some equity in the two AI music firms, as they have done in previous licensing agreements with new tech companies, like Spotify.
Last summer, Sony, Universal and Warner came together to sue Suno and Udio for copyright infringement of their sound recordings “at an almost unimaginable scale” to train their AI music models. If these licensing deals were to go through, it would likely lead to a settlement of the lawsuits.
Universal Music, Warner Music and Sony Music are in talks with Udio and Suno to license their music to the artificial intelligence startups, Billboard has confirmed, in deals that could help settle blockbuster lawsuits over AI music.
A year after the labels filed billion-dollar copyright cases against Udio and Suno, all three majors are discussing deals in which they would collect fees and receive equity in return for allowing the startups to use music to train their AI models, according to sources with knowledge of the talks. Bloomberg first reported the news on Sunday (June 1).
If reached, such deals would help settle the litigation and establish an influential precedent for how AI companies pay artists and music companies going forward, according to the sources, who requested anonymity to discuss the talks freely.
Trending on Billboard
Such an agreement would mark an abrupt end to a dispute that each side has framed as an existential clash over the future of music. The labels say the startups have stolen music on an “unimaginable scale” to build their models and are “trampling the rights of copyright owners”; Suno and Udio argue back that the music giants are abusing intellectual property to crush upstart competition from firms they see as a “threat to their market share.”
Settlement talks are a common and continuous feature of almost any litigation and do not necessarily indicate that any kind of deal is imminent. It’s unclear how advanced such negotiations are, or what exactly each side would be getting. And striking an actual deal will require sorting out many complex and novel issues relating to brand-new technologies and business models.
Reps for all three majors declined to comment. Suno and Udio did not immediately return requests for comment. A rep for the RIAA, which helped coordinate the lawsuits, declined to comment.
If Suno and Udio do grant equity to the majors in an eventual settlement, it will call to mind the deals struck by Spotify in the late 2000s, in which the upstart technology company gave the music industry a partial ownership stake in return for business-critical content. Those deals turned out to be massively lucrative for the labels and helped Spotify grow into a streaming behemoth.
The cases against Udio and Suno are two of many lawsuits filed against AI firms by book authors, visual artists, newspaper publishers and other creative industries, who have argued AI companies are violating copyrights on a massive scale by using copyrighted works to train their models. AI firms argue that it’s legal fair use, transforming all those old works into “outputs” that are entirely new.
That trillion-dollar question remains unanswered in the courts, where many of the lawsuits, including those against Suno and Udio, are still in the earliest stages. But last month, the U.S. Copyright Office came out against the AI firms, releasing a report that said training was likely not fair use.
“Making commercial use of vast troves of copyrighted works to produce expressive content that competes with them in existing markets, especially where this is accomplished through illegal access, goes beyond established fair use boundaries,” the office wrote in the report.
Even with the legal landscape unsettled, some content companies have struck deals with AI firms. Just last week, the New York Times — which is actively litigating one of the copyright cases — struck a deal to license its editorial content to Amazon for AI training. Last fall, Microsoft signed a deal with HarperCollins to use the book publisher’s nonfiction works for AI model training.
Music companies have not struck any such sweeping deals, and instead have preferred more limited partnerships with tech companies for “ethical” AI tools. UMG signed a deal last summer with SoundLabs for an AI-powered voice tool for artists and another one in November with an AI music company called KLAY. Sony made an early-stage investment in March in a licensed AI platform called Vermillio.
Stockholm-born music library giant Epidemic Sound is launching a new remix series, called Extra Version, on Wednesday (May 28) with help from DJ/producer Honey Dijon. As part of Extra Version, Epidemic pays participating DJs and producers “five to six figure sums” to pick from the songs, stems, samples and loops in its catalog of over 250,000 pieces of IP — and remix them into something new.”
Epidemic then adds the results to its ever-growing catalog available for use by clients — like content creators, advertisers and brands looking for easy-to-clear songs to soundtrack videos — and distributes them to streaming services.
To kick it off, Honey Dijon flipped the Epidemic-owned song “Umbélé” by electronic artist Ooyy and Swedish Grammy-award winning performer Ebo Krdum. “Teaming up with Epidemic Sound was a vibe,” she said in a press release. “They’re shaking things up in the best way… It’s all about freedom, fun and keeping the groove 100%.” The company plans to also collaborate with Major Lazer co-founder Switch and rising Korean talent Jeonghyeon on future Extra Version editions.
Trending on Billboard
With this series, Epidemic Sound CEO/co-founder Oscar Höglund tells Billboard he wants to show off the “high quality” of Epidemic’s catalog, which he believes rivals the quality of traditional major label releases. “The art of consuming music is changing,” he says. “It’s going from being a spectator sport to being a participative one. People want to remix their favorite music, they want to collaborate, and they want to create.” Epidemic, he continues “is creating the opportunity for incredibly talented producers and remixers and DJs to collaborate [and remix] our catalog. And then, we will help them distribute their remixes around the globe [both to streaming services and to their platform which provides pre-cleared music to content creators] – and they’ll get paid well while doing it.”
In this interview, Höglund talks the ins and outs of Extra Version, the ways he is integrating AI remix features to “create more use cases for the same songs,” and why he feels the allegations that Epidemic Sound has filled Spotify mood playlists with “ghost artists” is “deeply offensive to the artist in question.”
Why are you launching Extra Version?
We’ve seen that even though culture is moving towards [more participation and remixing], creatives have held back from doing it because from a legal perspective, it’s very hard to get rights to music because of fractional ownership. Our catalog has been built up, for almost a decade and a half, around the premise of having all the rights in one place. There’s nothing fractional about it. We own all of our music, and we are more than happy to offer the catalog up to producers. We just want to do it in a way that works for the artists who originally made the music [that Epidemic now owns], the remixers, the creators and the platforms where this music will go live and proliferate.
What is the payment model for Extra Version participants, and how does that differ from how producers are typically paid for remixes?
What often ends up happening is that producers are asked to create a remix for an artist, and they don’t get paid much to do it. Rather, the logic has been, the more culturally relevant the artist in question is, the lower your compensation, because your payment is in the cultural value you receive as a remixer from being associated with the artist.
We took a contrarian view here. We’ve always prided ourselves on putting our money where our mouth is, so instead we’re paying much more handsomely up front [to Extra Version participants]. We can’t disclose exactly how much, but I would say between five and six figures [for each remix]. We’re paying a lot up front, and this is not recoupable. It’s not a loan. It’s something the producer gets to keep.
Next step is that we allow the remixer to choose whatever track they feel creatively inclined to use from our catalog. Allowing for choice is a huge part of this. Remixers don’t have to license anything or worry about different samples being unlicensed. We own everything in perpetuity, and it’s all made available to you to pick and choose from. Then we will distribute the song. Most remixers don’t get a commission, just a flat fee. With Extra Version, we want to cut the remixers and producers in.
With Extra Version, Epidemic is opening up its catalog for remixes, and also making stems available so producers can mix and match the building blocks of your catalog. Does Epidemic have the goal of taking on sample or beat marketplaces like Splice or is this just for Extra Version?
When we started commissioning songs, we always got stems for everything. It is important from a soundtracking perspective. To the second part, the way we think about Extra Version is that this is not the end. We’re definitely not stopping here — rather, we’re saying this is the first step in our endeavor to help more music creators sustain themselves and democratize access to music. The bigger picture here is we want to help soundtrack the entire creator economy, and as such, we need to unlock our music.
So it sounds like the future of Epidemic Sound is offering samples, beats, and individual elements of the songs in the catalog to everyone, not just fully formed songs?
Correct.
Do you see Extra Version as an ongoing series, or is it a limited run?
This is not a limited run series. It’s the starting point of ushering in a completely new paradigm, one which is much more centered around the remixing and collaborative nature of culture. This is something that we’re deeply committed to and we’re going to spend a lot of time experimenting and seeing how this space is going to evolve.
It feels like you’re moving in the opposite direction of major labels. Nowadays, competitive deals at majors regularly involve the artist getting their masters back eventually, but the advance the artist gets up front is recoupable. Meanwhile, Epidemic is asking for full ownership of an artist’s tracks, but you provide non-recoupable money up front for the song. What is your goal with this approach?
At our core, we’ve taken one fundamentally contrarian belief: if you’re an artist, common wisdom says you should hold on to all of your [intellectual property]. That’s the traditional music industry right now. We think, in order to provide wide distribution and to provide superior monetization, we need to own 100% of the copyright. If we can build a platform, like we have, where there’s one point of contact when you want to license the song, then we can indemnify our customers and allow them to use the songs across all platforms, in all jurisdictions, and in all different scenarios. This allows for us to create predictability with Epidemic, so we can also pay our artists more predictable fees, too.
How is Epidemic thinking about AI?
We think that AI is an incredible tool to help augment human creativity, but never replace it. We’ve so far found that there’s tremendous amounts of value in using AI to help both music creators and video creators. We can use AI during the recommendation phase. If you’re a video creator we can use advanced AI search tools to help recommend tracks.
The old paradigm was, if you found a track that you like, suddenly you had to spend hours trying to re-edit that track such that it perfectly fit the video story you’re trying to tell, often with huge challenges from a legal perspective. “Am I even allowed to change the composition of this track?” The answer is often no. We’ve now been able to use AI [so] that, if you are a content creator, and you find this one track that you want to use to soundtrack your video, you can now speed it up, slow it down, change it. You can cut it. You can edit it — not replace it. This helps create more use cases for the same songs. Where there might have previously been 10 content creators who can use your track, now with adaptation maybe 20 or 30 or 100 creators will use it. That means the track is going to get played more and it’s going to earn more royalties [on streaming services]. And so ultimately, the human who made that track is going to make much more money, because AI has augmented the use cases.
Point three is purely generative — we’ve launched a product called AI Voice. We’ve gone to human voice actors, and we’ve struck agreements with them such that we pay them up front, we train and we use their voice, and then we allow our customers to use their voices. Every time they do, there’s an additional royalty so that the voice artists make additional money. We also put their personal emails out there in case content creators want to work directly with them. So suddenly, even when we go into the generative world of voice, we’re seeing that voice actors get used more, get more work, and get paid.
There have been allegations dating back to 2016 that Epidemic Sound has a deal of some kind with Spotify to fill some Spotify playlists with royalty-free music. It has been highly criticized. Can you explain what that arrangement is, if there is one?
I’d be happy to. Epidemic parallel publishes all of its music to all of the major DSPs around the entire world. We do that for a couple of different reasons, but the primary reason is it’s in our artists’ best interest, because we realized early on there was a Stranger Things, Kate Bush effect, meaning when Kate Bush’s track was used in Stranger Things, there was a massive surge in that song on streaming platforms around the world. We realized early on that that happens [when content creators use our songs].
There was also an adjacent trend, which we also tapped into very early, on streaming platforms in general — Spotify being one of them — that there was much more lean-back listening going on. The role of the record [or album] as the [driver] for music consumption started to diminish. More and more, [people were listening to] standalone tracks, but then ultimately playlists started to proliferate and come into their own.
Many of [the playlists] are hits-oriented, but there’s a huge proportion of playlists which are more functionally oriented. We do incredibly well across all the different playlists where people are looking to get to a specific theme or in a specific emotion. There’s music for sleeping, for concentrating, for studying, for getting ready, for meditating or for walking your dogs. Because what we do at our core is soundtracking, it turns out we were really, really good at [those playlists]. And while other people were trying to get into the bigger playlists to create the hits of tomorrow, we just kept doing the thing that we do really well. There was huge demand across all different DSPs, so we started to grow. And we became very, very significant and very, very successful.
Do you feel that it harms non-Epidemic artists in these genres to be competing for spots on mood-based playlists with your music?
No. If anything, the contrary: I think the old articles about Epidemic artists are deeply unfair. There was speculation: “Who are these artists? Do they even exist?” They are super talented artists in their own right. I took issue very much with that.
Various writers have referred to your music that is on Spotify playlists as “ghost artists” or fake artists.” How do you feel about those titles?
I think it’s deeply offensive for the artist in question. If you are an actor, you can play multiple different roles because you portray many different characters. That’s second nature. Artists, like actors, have the right to express their creativity in a multitude of different ways. It’s always them who determines if they want to publish their music under one name or not. Odds are that their fans might think they are all over the place, so quite often what we see happen is that artists have one brand for a certain genre of music, and different brand for another kind. If you look at Elton John, Madonna they [use] aliases. I seriously doubt that Madonna and Elton John would like to be called fake artists or ghost artists. That’s them creatively expressing through a different persona.