Can we feel as strongly about AI music as we do for AI visuals?
We are at an important moment in the human relationship with AI. In the world of visual media, we’re starting to see a strong negative reaction to purely AI-generated images and video, especially when they depict people. For me, when I see a social post featuring an AI-generated human, I now feel a kind of anger, and I’m not alone. The comment sections are often filled with pleas for the creator to stop making this kind of content (and you might find me there from time to time…).
Apologies for the AI image, but it’s here to make a point (and no, I do not like it!).
The challenge with music and audio is that for most people, it’s not easy, or even possible, to tell when something has been generated entirely by AI. What I keep wondering is, if it were easier to recognize, would we start to react to AI music the same way we do to AI visuals imitating humans? We can’t know for sure, but the trend with visuals suggests it’s possible, particularly in music that features the human voice or expressive solo instruments, where the sense of humanity is arguably strongest.
Our relationship with AI in creative expression may ultimately depend less on what it produces and more on whether we recognize it as artificial. With images, that recognition can be immediate, and for many of us, unsettling. With music, the line is harder to perceive, so the emotional illusion can hold. But if AI-made music became as easy to identify as an AI face, I suspect our reaction may very well shift too. We might find it harder to feel moved, not because the notes are wrong and the mix isn’t perfect, but because the source is revealed. Awareness itself can dissolve illusion.
That’s what I think. What do you think?
Jeff, I am sorry for the slow response as I have just returned from business travel. I agree with absolutely everything you say here. Like you, I have been confident in my ability to recognize the traces of generated music (remember how we would play “name the synth” when listening to electro pop music?), but I expect I am being fooled a bit more than I would prefer to admit at this point…
Thanks for the comment, Jonathan. I do hope you are right, but as Jeff Rona and I are discussing in another reply, even my own ability to recognize the fingerprints of generated music is being challenged as the tech improves.
Hi Will, it makes me sad to hear of your experience. I hope that supportive tech that will help us all recognize generated music and make our own informed choices advances quickly- my team are among many that are working on this.