How Will AI Impact The Future of Music?

AI is a hot topic in so many industries right now, but in music or the arts in general, it’s an especially sensitive conversation. Some people fear that they’ll lose their jobs, others worry that the value of human creativity is going to be diminished even more than it already has been, and many are concerned that they’ll be left behind as technology once again makes entire business models obsolete. Plus, there’s stuff like this happening that is just… slightly alarming.

At this point, it’s clearly inevitable that AI is going to have an impact on the future of music, but what does that look like? We teamed up with Mike Epstein, the EVP of Digital Marketing at Cinematic Music, to get some answers. After seeing some of his Twitter “check-in” threads about different topics that are affecting the music landscape in 2023, we decided to tap in with some AI experts, industry professionals, and creatives in music to gain perspective, share information, and raise some important questions. What we ended up with was way too long to fit into a Twitter thread, so here we are.—Jacob Moore, Founder and GM of Pigeons & Planes

Cherie Hu – Founder, Water & Music

Creative AI may be the most disruptive technology for the music business since the Napster era of piracy. Already in 2023, over 10 different music AI models have been released by independent researchers and big-tech companies like Google and ByteDance, allowing users to generate custom tracks in seconds using a mere text prompt. Thanks to off-the-shelf music AI tools like Boomy, hundreds of thousands of AI-generated songs are now listed on streaming services. At large, generative AI tools for text, and visual art like ChatGPT and Midjourney have picked up tens of millions of users, forcing us to rethink traditional notions of creativity, ownership, and attribution.

The concept at the core of why creative AI could be so disruptive is that the friction between having an idea, and manifesting that idea in the form of a creative work, has essentially gone down to zero. Barriers in distributing music have already been collapsing, with anyone now able to list their music on streaming services for free. To see the same barriers collapsing in the creative process means that the playing field of who can be an “artist,” in the sense of being an active participant in the music economy, will continue to increase dramatically.

The concept at the core of why creative AI could be so disruptive is that the friction between having an idea, and manifesting that idea in the form of a creative work, has essentially gone down to zero.

Traditional music-industry notions of authorship and attribution will likely need to be thrown out the window with creative AI, as what is happening under the hood with these AI models is much more complex than directly sampling a song. The legal and regulatory landscape around generative AI is still so unpredictable that trying to make any future predictions right now is a futile effort.

That said, any historical conflict between music and tech companies has usually boiled down to stubborn information silos between artists/rights holders and tech developers, and the lack of a shared language between those two parties around what counts as “success,” and what industry issues matter and why. So, my primary advice to artists and music-industry professionals is to break down these information silos now while we’re still early, and start talking to and collaborating with music and audio AI developers to develop a culture of transparent dialogue.

This goes the other way as well—the best music AI developers are building iterative feedback and collaboration cycles with artists into their day-to-day workflows. Novel approaches to licensing, monetization, and brand protection that are being experimented with right now with AI models prioritize direct artist consent and collaboration with developers as an ethical default (e.g. Spawning).

As these music AI models continue to mature rapidly and enter the hands of mainstream consumers, I think now is the time for music professionals to jump in, start experimenting, and have their say on how these tools can grow and evolve.

Rob Abelow – Founder, Where Music’s Going

Generative AI is coming for music, whether we like it or not. Much like every other creative industry. Here are a few areas of impact:

1. There will be a flood of mediocre, indistinguishable, unoriginal music. Emphasis on context- and mood-based music, ambient, sleep, lo-fi, instrumental. ‘Good enough’ music, which has become plentiful in the last 5-10 years, will now go from the long-tail to the generative AI.

2. Music libraries and production houses will see the most change. The mass music-as-a-service companies will almost entirely transition to AI, wiping out a huge area for 9-5 musicians to earn a living.

3. We may see a paradox movement back to an artist and originality focus for listener habits. Artists over songs. You must stand out. You must connect. Authenticity & originality as a premium.

4. Artists who most effectively use AI as part of their creative stack will ride a new wave. That’s not just in the music, but in video, art, and more.

The takeaway is to learn and adapt. Know who you are. The era of ‘creating for the algorithm’ may come to an end. How can you beat an advanced AI at that?

Justice Baiden – Co-founder and Head of A&R, LVRN

What are your thoughts or concerns around AI and music for the coming months and years? 
Just like anything in life, moderation is key. I think AI will help a lot, but could hinder creativity based on that becoming people’s only tool.

What do you expect to happen if there is no moderation with AI?
With AI specifically, creativity could take a backseat, right? For me, AI is a tool and it’s supposed to be able to help you get from point A to point Z more efficiently. Creatively, AI can allow us to expand on ideas and even connect with people that can take it to a different space. But to be relying on it… like I’m even seeing things where people are getting AI to write them songs, but there’s still a percentage of human touch that’s needed. And that usually shows through an error. Computers try to be as accurate as possible. But as humans, we still have a percentage of error. And that’s a certain feeling we’ll lose if we lean too far. I feel like we’ll lose things that keep creativity sacred if we use AI as a crutch.

Right, it raises questions about things like work ethic  and authenticity. Those things could potentially be questionable, and you might start to question how artists are making music and if it’s even them actually making it.
Yeah, exactly. For me, it’s not even the question of how artists make music, it’s more so: are you giving your best effort? The only way we can really push for the best creativity is through effort and by exploration. Once that’s removed and there’s a crutch, it slows down the innovation and the creativity.

The creatives can get lost because they don’t understand how to navigate tech, and the tech people don’t necessarily understand music language. So I think they both need each other, and tech is not going to stop moving forward. But creativity can.

Technology is never going to stop moving forward, and AI is just innovation in technology. But in the music industry or anything in the arts, the language between creativity and technology hasn’t been refined. The creatives can get lost because they don’t understand how to navigate tech, and the tech people don’t necessarily understand music language. So I think they both need each other, and tech is not going to stop moving forward. But creativity can.

With AI and how it might impact streaming streaming services, you mentioned the idea of how AI will double down on this idea of targeting your fans. Can you talk a little about that?
Yeah, even beyond music, I think the future of community building is going to involve a lot of niche focus. Right now, everything is so fragmented, but it’s not going to stop being fragmented because as more people get access to the internet, you get more pockets of different subcultures. They all continue to build and get stronger. Creating your community from a niche perspective is going to be the only way to sustain.

Someone like Beyoncé, for example, makes things an event. She’s loud enough and big enough and has enough reach to get the world to pay attention and focus on one thing for at least a day, right? Newer creatives aren’t going to have that luxury and that ability to get the whole world to be on the same page for a second.

I was at a conference a couple of weeks ago and Mr. Beast, the YouTube sensation, was giving a talk. And I knew who he was, but there were so many people around me that were asking, “Who is this guy?” That’s crazy, right? He has the biggest footprint on YouTube, but the fact that people didn’t even know who he is allows me to understand that it’s just all about niches and it’s all about how big you can make your niche. But for everybody to know one person? I think we’re just getting further and further away from that.

There is no there’s no limit on niches though. I look at how AI uses different keywords to advertise or get new fans, because at the end of the day interests are interests and we’re influenced by the people around us. I could be walking down the street, and I can almost guess who listens to Tyler, The Creator and who doesn’t. I’m not going to be 100%, but I know that I’ll be over 50%. If somebody looks at me, I don’t think they’ll guess I’m a huge Bon Iver fan.

What I’m trying to say is that there’s no limit to how big your niche gets. I think Mr. Beast proves that. He can have a hundred million people watch his video in two days. It’s still a niche, but it’s an international niche. Back in the day it was terrestrial radio and TV and we all had to pay attention to one thing, but it’s just not the same anymore. So you can try to build your community and grow it as much as possible, but the days of the whole world stopping to watch a music video are over.

Yeah, it’s fascinating, right? Working  at a label and doing digital marketing, that’s something that we’re challenged with. How do you create that visual brand when people aren’t watching music videos in the same way as they used to​​​​​​​? Everything is transitioning to short form, and how do you tell that story if people’s attention spans are diminishing?
Yeah, I challenge that theory. I think short form content is the way, but there’s still an audience that likes black and white movies. There’s still an audience that would appreciate short films. It might not be explosive numbers, but if the product is good and the quality’s there, you start to train your niche to understand and go along with what’s important to you. I think human behavior is just training. We have been trained to enjoy short form content. Long form content just takes longer but you have to train, to be patient, and the content has to be good. But if you can do that, I think you build more quality fans, because they appreciate what it is that you’re building.

Ashley Maass, Creative Director

What types of questions do you think AI raises when it comes to music creation, distribution, and marketing? 
The recent advancements in AI technology have prompted a fascinating inquiry into its impact on the creative pursuits of musicians, particularly with regard to the authenticity of expressing genuine emotions through AI-generated lyrics. There is a debate on whether AI helps or hinders the creative process, and whether it leads to more emotionless music. There’s also the question of whether AI makes music production more accessible to individuals without specialized knowledge in music composition. It’s understandable that some artists have concerns about AI potentially replacing their artistic abilities, given that many individuals make a living through various creative fields such as songwriting, singing, producing, and mixing. Ultimately, the impact of AI on music creation remains an open question that warrants further exploration.

How do you foresee AI impacting the way artists create their music and the way fans consume it? 
Songwriting is often a deeply personal and introspective process that can be hindered by writer’s block. In such cases, AI-generated prompts can serve as useful aids to help songwriters overcome creative hurdles and stimulate new ideas. However, if AI is solely responsible for generating lyrics, the resulting music may be perceived as lacking in emotional depth and authenticity. While some argue that AI-generated lyrics may result in more formulaic and emotionless music, others contend that AI can help songwriters more accurately express their intended feelings and emotions, which they may have struggled to find the right wording for.

Tao Romera, COO of Soundraw

Share your thoughts, questions, or concerns around AI and music for the coming months and years. 
I think there’s going to be a shift in the creator industry. “Creators” are going to become more like “directors,” taking part in the whole process of creating a piece of work instead of specializing in one particular aspect. For example, until the emergence of AI, you had copywriters, illustrators, voiceover artists and music producers, to name a few of those specialized professionals. Thanks to AI, creators are going to be able to leverage a set of tools in order to create a piece of work. Specialized professionals are still going to be there because the AI tools and the new “directors”’ skills in each field will never be comparable, but it will allow a huge number of people to come up with pieces of work they would have never had the resources to create were it not thanks to the AI tools.

What types of questions do you think AI raises when it comes to music creation, distribution, and marketing? 
In my opinion, the biggest question is about copyright infringement and how to properly compensate artists for their work in case their music has been used to train an AI model. It’s terrible that some companies are trying to make money on the back of artists by using their work to train an AI model and then sell works created by that AI.It’s terrible that some companies are trying to make money on the back of artists by using their work to train an AI model and then sell works created by that AI.

It’s terrible that some companies are trying to make money on the back of artists by using their work to train an AI model and then sell works created by that AI.

Soundraw’s AI engine has been trained exclusively on music generated in-house by our music producers, so the music generated by Soundraw is free from any copyright infringements. We respect the artists’ work.

How do you think AI will impact streaming services? 
Depending on the genre, I think we will see a lot of AI-generated music occupying spaces in the streaming industry. For example, lo-fi music is a good candidate. But for music you want to properly listen to, as opposed to music you play in the background while you do something else, I don’t think AI will become relevant.

How do you foresee AI impacting the way artists create their music and the way fans consume it? 
AI can actually be a great tool for artists, especially aspiring artists—people who can’t afford expensive beat lease licenses or a music producer/sound engineer. So rather than damaging the artists and the music industry, I think it’s going to have a very positive impact on it by allowing many more people to become artists and democratizing music creation.

Darryl Milliner – Artist Manager, SinceThe80s

AI technology provides many helpful applications, such as systems being used for composition through machine learning which can generate musical ideas or fragments that can then be refined by a human touch. There are also tools like LANDR which use AI to provide mastering services at affordable rates, a much more cost-efficient alternative to purchasing the necessary recording hardware and mastering software. AI algorithms are being used to create playlists on streaming services like Spotify, Apple Music, and Pandora.

AI tools are also being used by artists to create music, sounds, and references, as well as content. Content is vital for an artist to promote themselves, and AI content and art generators like Jasper.ai and Open.ai’s Dall-E are efficient and accurate systems. These programs are beneficial to artists and creatives as they save time, money, and keep the vision moving forward. Although AI technology has been used for many things in the music industry, I am still searching for a software that can take an uploaded song or sound and generate a piece of cover art or a visualizer based on the audio rather than a typed summary in a field box.

I would also like to mention that I used Playground AI to help me properly edit and get my ideas across more concisely in this letter.

Concerns:
My concerns regarding AI are that this software could begin to be used in a way that promotes full replacement of any human element or touch in this process. Creating music, marketing plans, and rollouts require cohesiveness but also a human experience. Finding ways to connect with fans, family, and friends through music is the one of the most important factors of why we all love and listen to music. It influences and inspires us with a feeling of love and compassion. I am not sure a machine can ever truly learn that, no matter what level of sophistication.

I have a concern that people will begin relying too much on this technology rather than going fully through the ups and downs of a creative process, which is usually what can be felt through the works once released. Keeping the human element and touch is essential and hopefully will never diminish.

Stefan Heinrich, CEO of mayk.it

Disclaimer: Opinion focused on music creation in particular, not about other ways AI can help artists.

I see the evolution of music and how it’s created and consumed as very similar to the evolution of the video space: Previously there was just Hollywood and PGC (100% access gated), then YouTube started and a few more people could create (via prosumer video editor, 10% access), then TikTok came out and way more creative people got access to simple tools to create videos (25% access). AI-generated music now will move access up to 90%.

AI will enable artists to share a much higher quantity of songs, and then have an increased chance to win in the sea/avalanche of mass UGC and AGC music that is about to flood the gates.

As we see with video and entertainment, Hollywood (PGC), Creator Videos (UGC) and now AGC (AI-curated content, e.g. Stable Diffusion) can co-exist, given lots of people have different tastes that are changing generationally. Similarly, professional music (PGM, labels, indie artists), creator music (UGM, e.g. mayk.it) and now AI-generated music (AGM) will co-exist. The timeline to move from PGC to AGC has been accelerated compared to video, where we almost skip the TikTok phase. When you flood the world with new stories and voices, now anyone can find stories that represent them, which wasn’t possible before for literally any niche.

Another important behavioral change that has already happened in the TikTok era is the commoditization of content. The game is mainly about quantity now—if you pump out 100 videos or songs a week, there is a good chance that one of them is going to be more successful and find its audience. In music we have been stuck in the library phase (Spotify, SoundCloud era = create perfect songs to last) and just started moving into this era last year. AI will help accelerate artist wins.

So will AI music replace musicians? Definitely not. It will enhance their workflow and help them focus on the most creative part of the song creation process: the core idea and feeling of the song. AI will enable artists to share (publish) a much higher quantity of songs to social media given the AI-assistance and then have an increased chance to win, given their creative capabilities, in the sea/avalanche of mass UGC and AGC music that is about to flood the gates. The best creative spark and relatable idea will continue to win. More artists can just play more and have more chances now.