AI in Music: Opportunity or Threat for Creators?

If the last century saw humanity experiencing the effects of the so-called arms race, we will remember this period (and perhaps the next one as well) as the era of the algorithm race. In recent years, the development and implementation of Artificial Intelligence across the entire user experience ecosystem have relentlessly pushed in the direction (or drift) that social media has chosen to take: keeping visitors’ eyes (and thumbs) glued to the screen for as long as possible.

Their strategy: emotional reinforcement. For example, if today I decide to listen to a lo-fi music playlist on YouTube, maybe just by chance, perhaps because I liked the cover image, tomorrow it will suggest similar videos. The more I listen, the more my feed will be saturated with them. If I happen to like a few videos of American highway patrol officers, I will soon be bombarded with escalation episodes, amateur denunciation videos, and so on. Sooner or later, surrounded only by this type of content, my mind will reinforce the association between police and violence, and rather than believing in a multifaceted world, I will convince myself that the world is deeply polarized—because that’s what I see constantly, so it must be true (at least for our monkey-evolved brain).

Playing on emotional responses is often highly profitable, just like shock culture in the art world. This explains why YouTube Music recently introduced "Ask for Music," an experimental AI-based feature that easily creates playlists based on user prompts.

“What are you in the mood for?” is the slogan of a trend that, in the world of music, has led to a proliferation of artificially generated content, often characterized by mass production and a standardized mediocrity — ironically, devoid of experimentation, surprises, or complex emotions.

The influence of AI — or rather, the way some people intend to use it — has impacted the entire artistic landscape (and beyond):

• In photography, it now competes significantly with traditional image creation.

• In news (and misinformation), it enables the falsification of audio, photos, and videos.

• In the film industry, where there are concerns that technology could threaten jobs in animation and sound design. Hollywood writers have already fought a major battle, among other things, to prevent studios from using ChatGPT to generate a 100,000-word story and then asking them to adapt it — so they can cut costs and lower wages (with the logic of: "It’s not your original story, you worked less, I’ll pay you less").

The social media game is changing — or rather, it has already changed — and in many ways, it is no longer sustainable in the long run. We constantly witness creator burnouts, and we only care to a certain extent because the void they leave when stepping back is quickly filled by someone ready to do anything to take a step forward, to have a moment of fame and redemption. Major social platforms have adapted to a system where subscribing to creators is almost unnecessary because their content will be suggested to users anyway in the "For You" pages designed specifically for them.

The underlying "non-problem" is that people like Ivan Linn — a pianist, entrepreneur, collaborator with major record labels (Sony, Warner, Universal), and member of the UNESCO Cultural Committee — see AI in music production as a way to revive traditional music and endangered sound cultures, as well as a tool to extend human expressive potential rather than replace it.

It must be admitted: as a creative tool, artificial intelligence holds great potential. This was demonstrated at Italy’s Sanremo Music Festival when they presented AI-generated cover images for participating songs, created by feeding the lyrics into a visual generator. The experiment wasn’t bad, but here’s my (personal) doubt: if they had asked real visual artists, would it have been a repeat of the scenario from a few years ago, when orchestra members were paid only 50 euros per night? Was this decision made for the AI hype, or to cut costs?

And so, here comes the first problem: there have always been people in charge of businesses who, when it comes to paying for creative work, fail to understand the value of a logo for their brand, a video, a photograph, an artistic concept, a song/jingle for their ad, or a website. They look for the cheapest solution (that’s what cousins have always been for), only to end up with poor results, which they then take as confirmation that these things aren’t worth investing in, that they don’t matter, that people don’t like them, and that spending money on them is pointless. However, in the past, such people would have been left with no content or low-quality, perhaps even trashy content. Today, and who knows for how long in the future, these people have found a way to be competitive without paying what is due to those who deserve it.

The second problem: even for those who willingly spent money in the past, capitalism demands constant economic growth. In recent years, layoffs and unpaid wages have been interpreted as economic progress. Right now, generative AI is in the hype phase, and its output is met with interest, but it is alarming to think that without regulations, no one (at present) can prevent AI from replacing even more jobs with these tools.

The issue is complex because it spans multiple domains (writing, visual arts, and music, to name a few). That’s why I set aside my perspective on visual arts and asked for insights from a group I recently had the pleasure of meeting — whom I appreciate for their musical style (and also because I play D&D with my friends). I’ll leave you with them and the interview we conducted:

1. Who are the Supernova Collective? How did the idea of creating music for RPGs come about?

Hi! We’re Matteo, Gianluca, and Giovanni — three composers and sound designers passionate about role-playing games. We started this journey together after meeting in a specialized course on music production for cinema and video games. We slowly began collaborating with various creators, including Italian groups that stream homebrew campaigns on Twitch and YouTube.
After a few months, we thought: why not create soundtracks for official Dungeons & Dragons manuals? And here we are.

2. What has been your experience trying to get your music noticed on platforms like YouTube?

It’s definitely challenging. We’ve experimented a lot with titles, thumbnails, descriptions, tags, and how we present our tracks.
On YouTube — and not just there — thousands of new pieces of content appear daily, and it’s not easy to get pushed by the algorithm and attract attention. The hardest part is convincing someone to click on your content... it makes a huge difference, sometimes even more than the quality of the music itself.
Of course, as views and comments start coming in, the content gains credibility for new users, and growth becomes almost automatic. But reaching that point is a long journey, especially in a competitive environment like this one.

3. What is the most frustrating part of seeing AI playlists gain so many views?

When we start working on an album, we study the manual, consult various masters, choose an artist to commission the cover from, and create a prospectus to decide on titles and descriptions. There is a lot of work behind it, not just our own.
It’s frustrating to go onto the platform and see 4-5 hour videos with AI-generated tracks simply lined up, presented with very evocative AI illustrations (a style that has now become the standard) that rack up more views than yours. We think this is the most frustrating part: seeing immediate results and the difficulty of keeping up with certain rhythms.

4. Do you ever feel tempted to use AI tools to speed up production or improve visibility?

Well, we’d be lying if we said no when it comes to illustrations.
It’s not easy to accept falling behind in terms of graphic possibilities; it would definitely be more functional (and economical) for us to generate a custom image for each video. We’ve noticed this works well for other creators.
In our early posts, we used two images generated with Midjourney. These were videos that weren’t part of specific albums, and we didn’t know who to turn to for graphic design. But we quickly realized it wasn’t the direction we wanted to take.

5. What strategies are you trying to adopt to get noticed?

We’re trying to focus on ease of use and content quality.
On YouTube, we upload looped versions of the tracks (30-60 min) so that there are no interruptions during the gaming session. We include titles that link to chapters in the manual to make it easier to find the masters. We create themed albums and playlists with the name of the official adventure.
As we grow, we’re learning what works and what doesn’t. Like with everything, we adjust along the way.
Musically, we try, of course, to give the tracks a recognizable style and stick as much as possible to moods that reflect specific situations. Otherwise, why would a listener prefer AI?

6. What do you think the music community (and the platforms themselves) should do to better support creators who don’t use AI tools?

We believe it’s not the responsibility of the music community or the platforms (from their point of view, AI content just does what’s convenient: get views).
The only way to address the issue (if it is an issue) would be to mandate that AI/non-AI content be properly and visibly flagged/tagged, limit monetization, and create specific search filters. It should come from “the top,” and it should be a global measure, but there are clearly other priorities at this historical moment, and there’s little awareness. It’s pretty utopian.

7. Some past technological changes, like the rise of home production for music or the advent of smartphone photography, have profoundly shaken their respective industries. Do you think the spread of AI in music creation is comparable to these revolutions? Or do you think AI represents a threat or an opportunity on a different scale?

Technology itself is never a problem. It’s a tool, a very powerful one, and like all powerful tools, it leads to new opportunities, revolutionizing the fields it’s applied to.
However, the problem we pose feels almost "philosophical."
We’ve spent centuries, decades, hoping that tedious and heavy jobs could be simplified to give us more time to dedicate to creating something beautiful. Something that is ours, that represents us, that makes us feel free to express ourselves.
Why hand over the entire artistic production chain to tools that are so far from being human? To do what? To get what?
We’re not opposed to AI as a tool; it would be utterly anachronistic. But we have to ask ourselves what we actually want to achieve.
What are we putting into this work of ours? Does it represent us?

8. On one hand, the accessibility of production tools has allowed many to emerge, while on the other hand, the spread of AI has generated fears related to the loss of visibility and opportunities. Considering AI as a tool whose use determines whether it is beneficial or harmful, what do you think should happen today? What are online platforms missing to protect human creativity? Are users and artists already making requests in this regard? If not, why isn’t action being taken? And finally, what could modern strategies for effective protest look like in the age of social media and the internet?

We want to hope that the awareness-raising effort will be natural and gradual. What’s missing is helping people understand how much work goes into it, what it takes to achieve results.
Probably the best way to do this is to get kids involved with art, with their passions.
AI is a relatively young tool; it’s normal that it’s under-regulated and that there’s so much curiosity about using it (almost bordering on abuse).
Maybe a first step to highlight the “problem” could be to start adding “NO AI” flags to productions, hoping that one day someone will take charge of the situation.