The old defense of AI-assisted art was simple: the human is still the author, the AI is just a very sophisticated brush. That argument is getting harder to hold. When Jason Allen won a Colorado art prize with a MidJourney image in 2022, the U.S. Copyright Office eventually declined to give him the copyright, citing the AI's 'inseparable contributions.' Not a bug in the process. A contribution. The legal system, which is usually the last institution to notice anything, spotted something real: at some point the tool stopped being a tool and started being a collaborator.
The Copyright Office's January 2025 report made it even blunter. A prompt, no matter how carefully constructed, isn't enough to make you the author of what comes out. The output goes straight to the public domain. Which raises an uncomfortable question: if a brilliant, specific, deeply intentional prompt still doesn't count as authorship, what exactly is the human doing? And if the answer is 'curating,' 'directing,' or 'selecting from outputs,' we should be honest that those are real creative acts. But they are not the same creative act as making something.
The Collaboration Nobody Agreed To
Here's where it gets genuinely strange. A LANDR survey of over 1,200 musicians found that 87% now use AI somewhere in their creative workflow. That's not a fringe phenomenon, that's the industry. Sixty million people used AI to make music in 2024 alone. At that scale, we're not talking about edge cases or provocateurs. We're talking about a fundamental shift in what making music actually involves, happening faster than anyone's vocabulary can track.
And yet audiences seem to know something is different, even if they can't articulate it legally or philosophically. In a BPI/AudienceNet survey, over 80% of UK music listeners said they value human-made music more than AI-generated music. They're not anti-technology. They want AI labelling (81.5%), they want government limits (71.5%). They're drawing a line, even if they'd struggle to explain exactly where it sits.
The Honest Version of This Question
The honest version of this debate isn't 'is AI art real art.' It's more unsettling than that. Creativity, as we've understood it, was tied to intention running all the way through the work. The choices, the constraints, the failures, the specific hand behind the thing. When a system can improvise within your style, finish your sentences with genuine fluency, and produce outputs you couldn't have predicted but immediately recognize as good, the intention gets distributed in ways we don't have clean frameworks for.
That's not an argument for banning these tools, or for pretending they don't change anything. It's an argument for being precise about what we're actually claiming when we say someone 'made' something with them. Directing an AI is a skill. Curating outputs is a skill. Neither is nothing. But calling it the same thing as authorship, without any asterisk, is fuzzy thinking dressed up as progressivism. The vocabulary needs to catch up, and the people building these tools have exactly zero incentive to help it do so.