I was recently asked by a friend what she should think about when it comes to music and AI, so I put a document together and decided to share it here. Although AI has been around a long time, it had a big consumer moment late last year with the launch of ChatGPT and the breakout uses of Midjourney and DALLE-3. All of these ‘programs’ are referred to as LLMs (Large Language Models).
The explosion in popularity of these LLMs has brought a lot of interest and excitement to the space. It forced Google to play its hand in its search evolution, launching Bard, which has become a dominant component to Google’s search engine and is steadily and slowly rolling out groundbreaking features. Plus, LLMs have brought excitement to the tech market with regards to investment, new start-ups, and new consumer use cases.
New announcements and apps are dropping daily, but to wrap your arms around the state of things right now, here are the areas to think about when it comes to music AI:
1. Fake Drake
This is not an LLM. There is no computer making music sound like Drake. Rather, someone recorded this song and created an “AI filter” that sounds like Drake. The same goes for Biggie, Sinatra, John Lennon, et al (they’re all amazing BTW and worth a listen). The press is calling this AI, but in truth it should be viewed like a TikTok filter. There’s an anonymous producer named ghostwriter creating some buzz with these “remixes.” He appears to be the breakout AI filter star to watch, but we’ll see if he has a leg to stand on.
There’s a strong push by the artists and labels for regulation. There’s no legal way these tracks should be allowed to use the artists’ names, but the question is, can they use their vocal sound, cadence, and style? It’s an important new conversation and my POV is as a deep fake it’s not legal, but as its own creation of art borrowing from other styles, it should be ok. Who’s in charge of regulating this matter remains unclear.
2. More Music
AI music has another seismic shift here, too. In an age where 100K songs are released per day on Spotify, what happens now with the click of a button? Infinitely more will drop. The streaming services are at risk of being flooded. The way artists are paid is already up in the air with many hoping for a shift to a more artist-centric streaming model that allows for cutting through the ‘filler tracks’ and paying the artist per stream on a per use basis. The long tail is going to get even longer and the models need to be written to protect the creators.
3. Mood Music
With the data collected by AI and the tools up to speed, you can create a song with any mood, length, or vibe just by using the right software. This is a game changer for background music in commercials. Musicians’ sync businesses are going to take a hit.
4. Elevation of the Bedroom Producer
Now that any noise can use voice filters, AI beat makers, harmonizers, lyric writers, or software to enhance their music, the aspirational musician is on a new level. Anybody can make a love poem. Karaoke is going to get better. Amateur producers and newcomers will produce with much more ease. Apps such as Audioshake allow you to separate stems of songs in ways we’ve never seen before. What does this do? Yes, it makes music even more ubiquitous, but also makes the demand for the best in the biz go up! This is worth a watch.
5. Curation > Creativity
In a world where there is a firehose level of demand for social content, curation is going to be more important than creativity. The demand in a social media era for more content has been exhausting musicians and creators. Consumers are being trained to want more and more, and creators who do so are being rewarded by the algorithm. AI is providing tools to creators that makes it easier to create. It’s an ugly world, but you need to know how to use the tools with your writing, video creation, and images.
6. Artist Cloning.
It’s better for artists to control their likeness and create their own clones. There are amazing new platforms like 1on1 that can pretty much turn you into a live chatbot of yourself. The company sits down with you, films you, learns your voice and how you would answer questions and then can put you in venues to have you directly talk to fans. Jerry Jones is doing it. Afterparty is also launching AI versions of artists to attend virtual events. This is the next evolution of chatbots, which had their breakout moment a few years back and are steadily being integrated into customer service all over the web.
7. Avatar Creators
Once Lil Miquela blew up, it became clear that avatars are going to be more and more viable as pop stars. These AI stars are easier to manage, crafted to perfection (even more so than K-pop stars), and scalable.
8. Streaming Services Enhancement
Streaming services are utilizing AI to enhance lyric integrations, create customized art, and even birth the AI DJ. Spotify has an AI DJ (called X) that knows your preferences and not only recommends music but also blends it and gives you personal insights. X is a cool feature but I think the name is going to have to change thanks to Elon. Perhaps more remarkable is Spotify’s new AI to clone voice tool that translates podcasts into other languages for global markets using the original speaker’s voice!
9. Beyond Songs (Rapid Fire)
There are so many ways that AI can and will disrupt the music business. There are reasons why top tech executives call this the most game-changing creation in human evolution since the ability to control fire. Just off the top, I’m seeing examples where AI is affecting:
- How music videos are being made
- Effectiveness in artwork (for social and album artwork)
- A&R decisions (finding acts)
- Management work (logistical work, especially)
- Efficiencies in contracts
- Efficiencies in tour routing
- AI Automatimated Marketing (SymphonyOS and UnitedMasters just joined forces to grow artist fan bases with automated marketing)
- Effectiveness in tracking song clearances
- Efficiencies in finding royalties and tracking unregistered song clearances
Here are a few artist use cases to think about:
Will.I.am created a productivity tool called FYI.AI. He totally understands the power of AI, is the ultimate artist advocate, and has access to some of the brightest minds in tech. This app is created for artists to project manage and uses AI in some of the best ways I’ve seen.
Grimes spun up Elf.Tech which allows her stems to be used for generative AI and is able to track the songs back to her for a shared royalty. It’s a great use of generative AI and also a reminder that when artists embrace tech very early it generally leads to a marketing boom for that act. We’ll see if other artists embrace it.
Lupe Fiasco‘s vision for 2073 in a music business AI world is beyond fun.
Brands are also embracing the new technology, partly because AI is hot and partly because it’s a fun way to be innovative. The intersection of culture, technology, and brand marketing is an expansive, collaborative field that is very buzzworthy. I’d expect more of this in the near future and heard our friends at Tuney.Io are cooking up some ideas.
What to look out for in my opinion is the following…
Ethics & Regulation: As things explode, there are a lot of important conversations around ethics and guard rails to protect humanity before ‘the robots take over.’ It sounds goofy but I think everyone understands it’s a necessary conversation.
Legal Battles: AI is limited to its inputs. Right now a lot of these LLMs are scraping and pulling from data streams to which they don’t have legal rights. The New York Times has stopped the use of any of their data for LLMs. This data is not free. Deals will need to be made to use all of these song lyrics, media sites, artists identities, and more. This is going to slow down the growth of this industry tremendously.
Huge Revenue Boom for Content Owners: Once deals are in place to leverage this content and data, this is going to be a massive new revenue generator for artists and labels. Artists and labels need to be cautious of AI stealing their content right now, but in the not-so-distant future there will be billions of dollars in new revenue.
Personal Data: The next frontier for LLMs is pulling from personal data. What if you can control the algorithm to only pull from your emails, pictures, and search queries? It’s going to know your taste even better than you know it yourself and create content for you on an even more sophisticated level. Imagine a world where I create an image or a poem for myself using thoughts like yours! We’ve been saying this for a while but in an AI world, data is more valuable than ever.
Individualism > Collectivism: This is happening in emerging culture already, and is a core element of American identity. Streetwear is no longer about luxury, it’s about personalization. AI is going to make this trend even easier to execute.
New Tools: New tools, features, and companies are debuting regularly. Moore’s Law is real and the technology is moving faster than we think. Yes, the big players in culture like Apple, Meta, Amazon, Microsoft, Nvidia, Netflix, Spotify, and Alphabet are better positioned than anyone to take advantage of this next revolution. We’ll see a world where the big get bigger, but there is also a chance for startups to build this new economy. Musicians are the ultimate influencers and can help lead the way. Music sets the trends and has since the start of the digital revolution. Music is right here leading the charge again.
Metaverse: Right now it’s still very expensive and high-touch to build in these worlds. It’s going to get easier and easier to create our own worlds using new game engines and technologies that will lead to continued growth here.
AI Music Governing Body: The music business is notoriously pretty fragmented but in this state of AI and ethics with laws and innovation all moving so fast, it’s pretty clear there should be a music industry task force around what’s happening. We need advocates for the industry as a whole.
AI as a Style like Y2K: There isn’t much data to support this, but I feel like this era of AI is going to lead to its own style in culture, from collage art to custom experiences, from new music to the way we dress, communicate, and feel. I believe this era will have a distinct style point that, in the nostalgia loop, we will reference 20 years from now as classic not unlike the revival of Y2K today.
We’re going to be talking about advanced marketing tactics in regards to AI and data at The Mondo Conference. Taking place the week of Oct 10th, it should be a great gathering of musicians and thought leaders. My talk is on Friday, October 13th at 2pm. If you’d like to check it out, here’s a 25% discount for a pass: beatsbytes25. There is a reason this one feels like CMJ. It’s because it was created by the founders of CMJ and takes place in Williamsburg. Come exchange ideas and talk shop!
Also published on Medium.