This new ongoing series is about the business of music with a SPIN twist.
Not since the dawn of the Internet, Napster, and streaming services combined has there been such curiosity and horror over the music industry’s latest challenge: AI. Every day brings a new headline pondering whether we are on the precipice of a new creative dawn or cascading into an impending downfall, as artificial intelligence seemingly outpaces human understanding.
In recent weeks, Universal Music went from suing AI music generator Udio for copyright infringement to partnering with it. Xania Monet hit a milestone as the first-ever AI-powered music artist to debut on the Billboard Adult R&B Airplay chart. And deceased artists like Don Williams have become targets of AI deep fakes with the release of faux “new albums.”
Despite different opinions about how AI may change the music industry, many people are rolling up their sleeves to figure out a way forward. Because the one thing we can all agree on is: AI is not going away.
One industry giant leading the charge is Charles Goldstuck, founder and managing partner of GoldState Music, a private investment firm focused on music rights. Through his myriad positions in the music industry, including as former president and COO of BMG, co-founder of J Records, executive vice president and general manager of Arista Records and executive vice president of Capitol Records, Goldstuck has been at the forefront of music publishing, artist development (working with everyone from Foo Fighters to Usher and Sarah McLachlan), and the advent of new technology. That includes helming TouchTunes Interactive Networks, where adaptive AI has become a very active part of its recommendation services.
Since 2023, Goldstuck has been involved in copious research about emerging technology’s latest twist—generative AI—and how it impacts everyone in the music ecosystem, especially creators. A graduate of the University of the Witwatersrand in Johannesburg, South Africa, Goldstuck has returned to this academic setting to fuel his research, launching the AI & African Music project at his alma mater while also working towards his PhD via the publication of four extensive white papers on the topic.
The first, titled “Past Precedent, Future Proof: Toward a New Legal and Commercial Framework for AI-Generated Music.” was published in August 2025 and takes the stance that if we don’t act with prudence, “the creative economy will be absorbed into an opaque, unregulated system in which human authorship is significantly devalued.”
As Goldstuck believes, and as was seen years ago with Napster and more recently with the Universal/Udio case study, legislation and lawsuits are not a progressive way forward. Instead, he suggests we need to develop a “multi-stakeholder” framework that protects creators while not dismissing innovation. Goldstuck’s thesis is that “the only mutually beneficial path forward is for the music industry and AI platforms to resort to negotiated settlements and collaboratively develop new licensing agreements.”
Here, Goldstuck shares more about his unique start in the industry, what the future of AI holds, and what music fans can do to protect artists.
You have a long résumé in the music industry, but what was the catalyst that inspired you to get started and eventually lead you to this important research today?
I grew up in a farming community in the middle of nowhere in South Africa. And at that time, there was a very restrictive government in place. The only way you could listen to music, if you lived in the country like I did, was on the radio because we had no television. Every language group in South Africa had their own channel. My twin brother and I were able to get a secondhand transistor radio and we would listen to these stations. So from a very young age, I was exposed to different sounds, African sounds, the Anglo-American repertoire, local Dutch-oriented repertoire, and I developed a very broad interest in music. When my brother and I got to university in Johannesburg, we had to figure out how to earn money to pay for our school fees. Part of what we did was booking bands on campus and getting involved in the music scene. But I always wanted to come to America because that was the epicenter.
What were you noticing about AI developments in 2023 that began steering your research?
The initial impetus was really a commercial one. It was after the launch of ChatGPT, now three years ago, and when I started seeing evidence of platforms like Suno and ElevenLabs raising significant sums of money to develop content creation platforms. The light went on for me because I had studied the YouTube phenomenon very carefully over the years, and being on the board of Audible Magic, I understood that fingerprinting and watermarking and traceability was so important in giving birth to user-generated content that could allow for UGC to be monetized. When I saw what was going on, and I saw the lawsuits being filed a little over two years ago, the question I kept asking was: How will generative AI impact the value of music? And what I realized was there was no research available.
What has your research unveiled about strategic ways we need to proactively move forward with AI?
In the first white paper, the approach has a legal and commercial framework, which in effect hypothesizes that rights holders, creators, and tech platforms are going to have to come to negotiated settlements and enter into licenses because the alternative is chaos that would not serve anybody well. And as we saw with Universal and Udio entering into a strategic pact, all these licenses are starting to flow through. But bear in mind, licenses today are very rudimentary. I do believe they will evolve in the same way as Audible Magic. When it started fingerprinting and tracing technologies, it was at a very basic level and then evolved. YouTube’s content ID is very sophisticated, but it also took time. My second paper is going to look at detection and attribution that actually work and allow for proper attribution, so that the rights holders get paid fairly and appropriately. I hope that by publishing all of this and treating it as open source information, as opposed to proprietary info within the walls of my investment fund, it’ll help educate others and help spur debate to encourage others to be collaborative in trying to develop the right outcomes.
Has your role as executive chair of TouchTunes given you more insight into the potential of AI?
As platforms like Spotify were using adaptive AI and machine learning to help develop recommendation engines, we took a page out of that book, where adaptive AI really became an enabler and a tool to allow us to offer a much better service, and it really made a significant difference in the user experience. So early on, I’ve been a proponent of using what people would think of as machine learning. I saw how powerful adaptive AI had been in helping me with TouchTunes, so I had an open mind because I’d been the beneficiary of what AI could do. But then very quickly I realized that there was no research to help guide us on where this would go, and this is what drove me to really stepping up my efforts to try and understand how to tackle this knowledge gap.
I think a lot of people, especially creators, are fearful of the full capability of this evolving technology and how it impacts their livelihoods. Is there any advice you have that may assuage that?
I don’t think you can avoid the fear because creators are, in my view, right in being fearful of what could happen. It is a rational and legitimate sentiment. But I think the dynamic really is, before you let fear take over and let that be the only reaction driving how you think about this, start educating yourself because there’s always a balance. If you take the songwriting community, I hear from a lot of songwriters, ‘We’re using AI to help us with our lyrics, to help prompt outcomes where we can improve what we’re writing.’ If you are somebody sitting in your home studio trying to write and record songs, you might use AI to import song structures, lyrics, sounds, and vocals in a way that you’d traditionally have to really go into a mainstream studio to do, so I think generative AI is just making that process so much more efficient. There are purists who will say, I’m not doing that. But I think by and large, especially with the younger and emerging artists, you’re going to find that the generative tools are going to be widely used. [Investment firm Houlihan Lokey, which included Goldstuck’s research in a “state of AI” report this year, found that the majority of music creators are already using, or are interested in using, AI tools; just 21% said they are not interested in exploring the technology.] But I think that the creative community has to be very open to understanding the benefits of generative AI, and there obviously has to be a concerted effort on the part of rights owners to be very aggressive in the face of very well-funded Gen AI content creation platforms. Because the Silicon Valley mindset is we’ll take what we want, come after us, and we’ll worry about it later. But by the time something is settled, they’ve raised billions of dollars and created a lot of value off people’s backs. So the industry is correct to be fearful, but also to be aggressive in enforcing our perceived rights as content holders.
One of the biggest issues right now is the influx of AI-generated content infiltrating streaming services, and it seems detection will be paramount in that realm. Do you agree?
Definitely. Deezer published white papers available for anyone to see where they estimate, as of last month, 28% of all uploads was generative or synthetic content, which they have taken down. With all this content cluttering up the streaming services, it creates the dynamic for a bad user experience. So I think they’re doing the right thing. Spotify indicated that they took down 75 million tracks that they didn’t believe should be on the network. Now, we don’t know which of those are generative content, but it’s going to be critical for the streaming services and these social networks that have music embedded in their offerings to be able to filter out synthetic content. Because if not, the service is just going to erode. And ultimately, I think it makes for a very bad dynamic in the creative community.
Many people have implored Congress to take more action with AI regulation. Do you agree that Congress should be doing more?
I split this into two constituents. The European Union, who passed the EU Artificial Intelligence Act 18 months ago and are currently eliciting comments for their code of activity. In the U.S., the Trump administration is less concerned about copyright infringement and protection of creator rights. They’re much more concerned, like the U.K., about America’s competitive positioning in the world relative to AI. So the balance is protection of intellectual property versus opening the markets so that our technology firms are not hamstrung by any regulation and can, in effect, develop offerings that are competitive versus the rest of the world. So because of that dynamic, you do see things like the ELVIS Act that was passed, and the NO FAKES Act. There have been efforts, but Congress generally doesn’t like to legislate unless it has to. So I don’t think you’re going to see Congress stepping in and setting up the guardrails, so to speak. I think the first phase of evolution will be the licensing regimes that are developed by rights holders driven by the major labels, and ultimately legislation might follow in the wake of those.
We’ve talked a lot about the music creator and the industry infrastructure. But in terms of music fans, are there things that fans should be aware of and maybe can do to help in this evolution?
I think it is about ultimately supporting your favorite artists who legitimately develop their music and are trying to make a living. Support your artists. Don’t jump on the synthetic bandwagon just because something like The Velvet Sundown attracted a lot of streams. That just takes away from legitimate artistry. We need to support artists, show up at their live events, stream their music. That is what the industry needs most.
