Real Future Of AI in Music: Production Copilot Coming To Your DAW?

Whenever Artificial Intelligence is discussed in the music production sphere, it’s almost exclusively about AI generated music/audio, and how they are worthless, soulless crap only good for YouTube videos or commercials. But I’m afraid that the majority of music makers have not realized what an LLM AI’s true strength is and what it’s meant to be - a productivity tool. The past month saw the introduction of Microsoft 365 Copilot and GitHub Copilot X, which I highly recommend checking out if you haven’t already, and get an idea of what a DAW copilot can look like in the future. It’s not about AI generating an entire wacky and wonky album, but assisting professionals to produce an album better and faster, as just one example.
Here is what GPT-4 has to say about it, and I personally found it fascinating. However many people might sneer at this concept, I hope you, the future facing Elektronaut, can find it at least somewhat inspiring:

In the future, a music production copilot AI could be a highly advanced and versatile tool that assists musicians, producers, and audio engineers throughout the entire process of creating music. Here are some potential capabilities that a music production copilot AI may possess:
1. Intelligent Composition Assistance: The AI could analyze a musician’s creative input, such as melody or chord progression ideas, and provide suggestions for harmonies, chord inversions, and melodic variations to enhance the composition.
2. Sound Design and Synthesis: The AI could generate unique and custom sounds using advanced sound synthesis techniques, and assist in designing sounds for different instruments, synthesizers, and virtual instruments. It could also analyze and identify sounds from audio samples, making it easier to recreate specific sounds or styles.
3. Arrangement and Mixing Assistance: The AI could help with arranging different sections of a song, optimizing the structure, and suggesting transitions. It could also analyze and optimize the mix, providing suggestions for EQ, compression, and other audio processing techniques to achieve a polished and professional sound.
4. Performance and Production Assistance: The AI could analyze and interpret audio recordings of performances, providing feedback on timing, pitch, and expression, and helping musicians improve their performances. It could also assist in the production process, such as automatically aligning and editing multiple takes, or automating repetitive tasks like vocal tuning or drum quantization.
5. Real-time Collaboration and Interactive Feedback: The AI could function as a real-time collaborator, providing instant feedback and suggestions during live recording sessions, jamming sessions, or online collaborations. It could adapt to different musical genres, styles, and preferences, and provide contextual feedback to enhance the creative process.
6. Music Theory and Education: The AI could serve as a comprehensive music theory and education resource, providing information, tutorials, and interactive lessons on various musical concepts, techniques, and genres. It could adapt its teaching approach to the user’s level of expertise and provide personalized guidance for learning and improving musical skills.
7. Creative Inspiration: The AI could generate new musical ideas, melodies, rhythms, and textures based on the user’s input, providing endless creative possibilities and helping musicians overcome creative blocks. It could also analyze and provide feedback on the emotional impact of the music, helping musicians evoke specific moods or emotions in their compositions.
Overall, a music production copilot AI of the future could be a powerful and versatile tool that enhances the creative process, streamlines production workflows, and provides valuable insights and assistance to musicians, producers, and audio engineers. It could revolutionize the way music is created and produced, opening up new creative horizons and possibilities.

7 Likes

That assistance would take me out of the loop of all these different processes.
Music is the only thing I put everything into.
I enjoy every aspect.
Without all those laborious sometimes boring moments I doubt I’d have ideas about how I needed to finish the piece the way I painstakingly battled through it.

I’m not giving AI my gratitude like that.
Steal my happy accidents, nah I’m good

25 Likes

If you are a producer for money not creativity and your only goal is to keep the wheel turning, I’m certain that kind of “professional” will always crave any way to move the process along quicker than the next guy. Will it be anything worth hearing or that keeps them in demand? How good is the AI? I don’t have the answers to questions that I don’t particularly feel comfortable even thinking about.

9 Likes

if you thought chain of title for copyright clearance was a pain in the ass before . . .

7 Likes

I’m probably a little close-minded here (I don’t even like using the automated mastering settings done by Izotope’s plugins and the like) but all AI does is detract from the creative process and its expression as a fundamental (the fundamental?) cornerstone of the human experience. this will be something that will make a lot of people a lot of money and I don’t think many musicians will be in that camp of people making money. and if there’s no money and no creative expression, why produce anything at all?

7 Likes

It’s just another tool that will increase your creativity. Those that use it right now actually say this, about their own process.

9 Likes

I was thinking something similar.
For work, if I used an AI rotoscope to speed up a shot getting complete, I’ll use it.

For personal creative projects, it goes against everything I’m in it for, and at the same time, give these algorithms some of the load I’d actually want to carry.

1 Like

The field of Artificial Intelligence is old:

  • The Lisp programming language is 63 years old and has been heavily used by AI researchers since it’s invention
  • The Turing Test itself is 73 years old
  • Automatons have been developed over the millennia, dating back at least to classical Egypt
  • Divination techniques likely predate written history and depending on your perspective are tools for communicating with the gods, or embody the gods themselves.

In other words, LLMs are new and cool and exciting, but AI itself is far from new. AI’s impact on music is probably also far from new.

Personally, I’m with @bodymechanics. I make music for fun. Dabbling with AI tools, like the I-Arp in the Jupiter X/Xm or generative tools like the NDLR is fun, but the ultimate goal is to keep the experience of making music fun. I enjoy learning, and playing with new tools and old techniques are both fun for me. But I’m not interested in a button to press that will generate a banger. I already own a few copies of Pretty Hate Machine and even know how to make Youtube play it on demand. :joy:

LLMs will likely make copyright clearance more transparent. Give Disney’s lawyers a chance to smash everyone with their giant mallets and provenance discovery tools will start to sprout like weeds.

Music is among the hardest ways to make money that I know of. AI tools might make that a little bit easier, but you can (allegedly) make absolutely insane amounts of money doing prompt engineering for AI companies now, so why not just do that?

For people who are absolutely committed to music-as-a-business, I can see a lot of opportunities for AI, but that just doesn’t put an idiotic grin on my face the way that playing with the filter on a 303 that goes into two sequential distortion units does. :upside_down_face:

8 Likes

I’m also just stubborn because it feels like we’re at our last stand with all this.
AI is going to take this world and completely rearrange it.
I’m really just going down fighting, because we’ll probably loose.

6 Likes

You and John Henry, the steel driving man.

2 Likes

So be it, I feel good about it

4 Likes

AI will (probably) be a great leap forward on the technical side – mixing, mastering, programming, theory tools for the untrained. Bring it all on

But to the extent that any system is relying on mining copyrighted materials at scale to become “intelligent” the owners of those protected works are going to come down on users like a ton of bricks. with sample clearance, at least it’s ascertainable what was sampled with or without permission. trying to explain how your AI-aided composition didn’t rely on the hundreds or thousands of copyrighted songs that might have gone into the program to come up with a “new” work is going to be a legal minefield, and highly expensive at that

1 Like

My point is that this fight is not new.

My opinion, is to work with it, and shape it to your creative passions.

6 Likes

If fighting brings you joy, by all means fight (in a non-violent way, I would hope).

I wouldn’t want to underestimate the stubbornness of humans, though. It has been a few decades that William Gibson observed that “the future is already here, it just isn’t evenly distributed”, or something along those lines.

The AI apocalypse narrative is almost entirely an Open AI marketing campaign. It works - we are talking about them right now instead of arguing about why we aren’t making tracks. :joy:

1 Like

They do need to fix the copyright issue, but that is fixable.

Excuse me for the extent i pushed this off-topic.

OP, guidance please so we stay on-topic.

1 Like

I’m into the aspects of using AI for something good.
I have little faith in us as humans to not rush into things without considering the consequences. To not go for the money.
It will rearrange us so much more than the last two decades of tech advancement.
I actually feel good going down as the old man yells at clouds. Just taking a stand for my last sense of self worth.

4 Likes

I’m unplugging my computer before I go to sleep tonight.

4 Likes

I’m not making tracks cuz I don’t feel like it.
The AI gold rush isn’t going to work me to the grace. I can do that on my own. :kissing_heart:

2 Likes

The thing to do is buy up all the “boring” '90s romplers. None of them have AI. Korg had AI^2, and that could be a flash-fad among tiktokkers, or whatever post-post-post-tok comes next.

4 Likes

The point I forgot to make when I was navel-gazing above:

A friendly buddy musician could be a cool AI tool.

At a minimum, it would listen to what you are playing, work out the genre, and jam along appropriately.

It would be nice if it could make positive critical suggestions that improve your composition and performance.

2 Likes