AI Video Edited by: Ian Rushton
Music Written, Produced, Recorded and Mixed in Cubase 14 by: Ian Rushton
Vocals: AI
Guitars: Ian Rushton and AI
Drums: AI
Keyboards: Ian Rushton
Bass: Shayne Williams
2 Likes
Ian, I have to admit, I found this one somewhat disturbing. It does sound good, though.
Just for starters, why did you use AI vocals, when you yourself are such a good vocalist? Possibly some issue Iām not aware of?
Same goes for AI guitar. You can play guitar!
And then the drums - Iām not familiar with AI tools for drums; how did you do that? Iāve been using NI Studio Drummer forever, programming every note. To the point where one of my live drummers transcribed it by ear so he could read it in live performances. Is that a thing of the past? How important was it that I wrote every drum note into midi?
Give us some perspective please! Iām not even going to comment on the video; not my field!
1 Like
Thanks for the thoughtful comment - and good questions. Itās definitely darker territory - so I get your ādisturbingā reaction. Interesting though, seems like that raw emotional intensity is whatās connecting with other people.
Calling me a āgood vocalistā is pretty generous! Thanks, but Iāve never thought of myself as more than OK on some songs. Guitar too - Iām pretty limited. Often takes me multiple takes and tons of editing just to get something passable. Thatās exactly why I used to bring other musicians in for online collaborations back in the day. Trying to be a jack of all trades always meant compromise, well for me anyway - my songs were always better with that outside input.
So whatās different now? Mostly access and affordability. Iām basically using AI tools in 2025/2026 the way I used online collaborators 20-odd years ago. Itās experimental for me - just seeing whatās possible when willing and able session players/collaborators arenāt so easily accessed anymore.
On drums - funny thing is, Iāve personally never written a single drum note! Iām simply not qualified. I mostly leaned on Jamstix with Steve Slate drums (or other), which was already pretty automated, claimed to be AI - even 10 years ago, but it does require a fair bit of fiddling regardless though. So this isnāt really a shift away from any kind of careful MIDI work, which I never did - more like the next step in the same direction I was already going years ago.
The basic process: AI does a ācover,ā I pull out the usable stems, mix them with my own parts. I could go into more detail but thatās the general gist of it. Iām just āAI-curiousā I guess and Iām actually having a blast exploring this - hearing and remixing old long forgotten songs in this new āreimaginedā way. And, the video side especially - total game-changer for me creatively.
Your decades programming every drum note? I wouldnāt call that I thing of the past - thatās your process and craft. This is just mine, which is evolving with these new tools available now to us DIY producers/musicians.
Thanks for explaining. The thing I like about your voice is it always sounds like you.
Iām not sure I understood about AI doing a cover. Is it that you submit a preliminary version, and it creates a better version, with stems?
Essentially yes - I just uploaded my old mixes - AI does a version (same tempo and key) and I then downloaded the stems and use select AI parts in combination with my original tracks to create a new human + AI version of the song. As it happens the free platform I was using has now disabled this feature - copyright concerns I believe. In my case of course I own the Ā© but I imagine others have likely been abusing the system by creating AI covers of copyrighted works they donāt own.
Anyway, a bit of a fun distraction - Iāll do a few more like this then must get back and finish of my new album āPensionerā !
Interesting. Seems inevitable to me that music composition by humans is doomed. And that AI music will never be able to get beyond what it was trained on. It simply does not have the capacity to innovate. But it does have the capacity to crowd out any human innovation. But that does not seem to limit its possibility to churn out huge amounts of music. So all the music we have now, maybe thatās it.
So Iām considering how much music I can contribute to the training pool before itās the final version of all human composition!
I see you as an original human contributor to the pool. What do you think? I saw that Bandcamp made a statement banning AI generated music, but I donāt see how they will be able to enforce it. Perhaps theyāll create an AI tool to detect it!
How will I know in the future that the music I love was made by humans? Or will it blur to the point where it canāt be determined?
I see myself as a composer/ songwriter/producer and musician - AI is just a new tool to help achieve the outcomes Iām after - sometimes.
Reality is, us DIY producers have been using virtual instruments, sound libraries, loops, VSTis, and all kinds of āartificialā elements for many years now. Where do we draw the line? Is a sampled orchestra more ārealā than AI vocals? Is a drum loop I didnāt play myself more legit than an AI-generated drum part? Weāve been in this grey area for decades I reckon.
Personally, I donāt buy the doom scenario. Human creativity and innovation arenāt going anywhere. Every major technological shift in music - synthesizers, drum machines, DAWs, auto-tune etc - people generally just accepted these innovations and integrated them into their creative workflows. Music composing and production etc didnāt end, it just evolved. To my mind, AI is just the next evolutionary leap in that progression.
I get thereās some legitimate questions about economic impact on say session musicians and studios, maybe demand for human-written jingles and sync and the like is under threat - but thatās wholly AI-generated works churned out en masse with little or no human creative direction - thatās a different conversation.
Interstingly, one thing Iāve observed: people talk about āAI slopā being soulless - but only after they know itās AI. The bias is in the knowing, not the listening.
I literally experienced that bias in action earlier today a it happens, with another track - one person (an accomplished guitarist) knew AI was involved in the generation of a certain guitar solo part in the song -thier comments were the usual soul-less, forgettable, no feel etc. Another talented musician contact who didnāt know it was AI generated said in total contrast: āI love the guitar work; it has a real jazz-tinged fluidity to itā. I then told him the part was actually AI-generated - āWow, I could not tell as the guitar piece had a real feel to it.ā
Nice having a discussion with you, since I admire everything youāve done on this forum. And I think Iāve been on this forum for as long as you have been on this forum. Youāre a legend!
For me, what it comes down to is whether music will move forward or not.
Imagine a new genre developing in New Orleans a hundred and twenty five years ago or so, rising from a tradition of marching bands, and spreading across the USA. How did that happen? I suppose it was players playing together, trying out the new tricks, figuring out how to adapt and modify what they considered cool.
Imagine some of their cool ideas were adapted by a NY Broadway composer. Imagine that what that composer wrote finds its way to France, and then imagine that a revered French composer starts using the same tricks, but adding his own new tricks.
This is all human beings playing together, listening to each other, finding what excites them the most, adding to it, and making new stuff.
I just donāt see any way that AI is going to be able to do anything more than re-do what already exists, in one iteration or another. Because almost all music we hear today is already re-doing what already exists. And AI will do it very well, and it will make most composers unnecessary. For games, for background music for documentaries and other things, for pop songs, even for film scores. No human composer will be able to compete, because the AI composers will be infinitely faster.
Out of this, there is no possible way to develop a new genre. No way to pass it from one musician to another. Once AI takes over and dominates everything we hear, music becomes frozen. To me, this is very sad.
But Iām going to keep composing as long as Iām alive.
1 Like
Yeahā¦ good to have a talk about this stuff - very topical these days!
Your thoughts clearly fall under the āwhole other conversationā I alluded previously - wholly AI-generated music churned out with minimal human input, versus using AI as a production tool alongside human creativity.
So, in that regard your jazz analogy is pretty much spot on for how innovation has historically happened - humans jamming together, stealing ideas from each other, adding their own twist, passing it along. That organic interchange is defintely real and valuable.
But hereās where I think differently: I donāt see AI replacing that process, I see it coexisting alongside it. Yeah, AI will probably dominate a lot of commercial/functional music - game soundtracks, jingles, background stuff etc as I mentioned previously, maybe even some pop. But does that stop passionate musicians from gathering, experimenting, and pushing boundaries? I donāt think so. That human drive to create something that excites us doesnāt need the absence of AI - it just needs humans who give a sh*t.
You say AI can only re-do what already exists. Fair point. At an individual level though, arenāt we all kinda trapped by our own influences anyway? - Iāve lost count of how many times I was told my productions sounded very 70ās or 80ā! 
I guess weāre all essentiallly recreating what weāve absorbed during those youthful years when we were most influenced by the music of the time.
And hereās the thing - a bit of irony that the song that kicked off this whole conversation is called āNothing New.ā 
Because Iām not sure thereās anywhere truly new left to go with music anyway. New genres have been just evolutions of what came before. Jazz came from marching bands and blues. Hip-hop sampled everything. Is there anywhere left to go thatās would be even remotely appealing to listen to? Is it possible AI could one day stumble onto completely novel combinations the same way humans did - that anyone would even want to listen to? I really dunno.
The fear that music becomes āfrozenā - I get it. But I think that only happens if we stop creating. And given how many bedroom producers, experimental musicians, and passionate weirdos (like me) are out there making stuff just because we have toā¦ with or without AI tools - well,Iām not personally worried about that.
Good discussion. Could AI stumble onto novel combinations that humans would appreciate? I sincerely doubt that. AI isnāt conscious, and doesnāt have emotions. Iām making an assumption that for humans, music triggers an emotional response. Which I think is the whole point of music in the first place.
Music can make you jump for joy and dance. Music can make you laugh. Music can make you sad. Music can generate awe. Music can make you feel a part of a family or a tribe. Music can inspire you. Music can make you wish. Music can make you wonder. And more. We donāt really know what music does to us. We just know we have to have it, make it, play it, listen to it.
Itās a language that can communicate things that words cannot, to paraphrase Esa-Pekka Salonen, the great Finnish composer and conductor. In other words, music passes emotions to an audience that cannot be expressed in language.
AI, on the other hand, canāt even hear the music it is creating. AI will have learned what is the most likely next section of the digital representation of the waveform for a song based on a prompt from words, having learned how to interpret all the words, based on everything everybody ever uploaded, together with having learned which words collectively are associated with which sequences of frequencies. Pretty spectacular, when you think about it.
I think AI could be programmed to make something novel, but it wouldnāt make any sense to us. And I think AI could be programmed to share with other AIs, and all of those AIs could make a lot of new music, but it wouldnāt do anything for humans. There would just be an AI top ten list.
None of these remarks contradict your point about AI as a useful tool. Iām only commenting on whether AI could create original music that would make a connection with humans, and I do realize there was an AI generated country hit recently, but what Iām saying is that there was nothing original about that hit. It was a programmed rehash. So AI can do that for sure, and thatās why Iām concerned that there wonāt be anything new, especially when all the people who make a living from creating music canāt make a living anymore.
Last point, I think what weāre seeing right now is all of the different people and companies who are trying to figure out how to make money off of what AI can do with music. They are certainly not trying to figure out how to advance the quality of music. And this is why someone making a living based on composing is probably doomed.
This is really enough for now. Is there a badge for longest post contributed?
1 Like
Hi Ian,
I really like the qualities of the AI voice that youāve created in this production. The feeling of a strained airiness in the vocal is right for the lyrics, helping to highlight the despair of the āsingerā.
1 Like
As it happens this rather timely Youtube video from David Bennett just dropped covering this very topic - he discusses many of the points made. Worth a watch I think:
Thanks for sharing this one. Heās making a lot of good points, that weāve touched on. I watch David Bennett from time to time. I think the one point he may be missing is that when the ability to make music for money goes away, who will be left trying to make new music? I mean even Schoenberg made his living making music!