Here's a paradox. The easier it becomes to create images, music and text with AI assistance, the more vicious fights are erupting over who actually created what. And I'm not that surprised. Because, honestly, I'm confused myself.
Earlier this month I got invited to Upscale Conference in Madrid, an event hosted by Freepik where I met a lot of creatives who are generating AI video and making money from it. Not using it for ideation or brainstorming. Actual, final content.
What protections you might have over your AI creations, though, is debatable. Notably, the US Copyright Office issued guidance this January: if an AI system creates a work entirely on its own, it can’t be copyrighted because it has no human author.
In other words, typing a short prompt into Midjourney doesn’t make you the creator. If you heavily edit or build on the AI output, that part might be protected, but only the human-made bits. Where that line sits is still very blurry.
The people making fast cash from generative AI right now, however, probably don't care all that much. The people who are really suffering? The people who've historically created the work these AI systems have been trained on… and who are increasingly finding themselves replaced by said systems.
Intrinsic value
This goes deeper than money, really. It’s about whether human creativity has any intrinsic value left. Or whether it’s now just convenient compost for machines to mulch through.
Right now, a thousand lawsuits are grinding their way through the courts while AI companies insist that hoovering up copyrighted work is simply “transformative fair use”. This argument boils down to, "Our machines are just learning like humans do.” But humans don’t gorge on millions of artworks in milliseconds, convert them into insanely complex maths then flog the results as a subscription service. If we could, we wouldn't be having this conversation.
And so we drift into a world where the line between an artist and their work is being smudged out, and artists themselves are increasingly disposable.
The Copyright Office is at least having a go at asserting that humans matter in all this. Nice sentiment. But it doesn’t solve the elephant tap-dancing through the room: that these models are built on mountains of creative labour nobody asked permission to use, let alone paid for (including, I should add, my own books).
So what’s emerging is a strange new creative underclass: people whose work is apparently priceless when scraped into a training dataset, yet mysteriously worthless when it comes to paying them. The easier it is for a machine to mimic them, the more disposable they become. And if that doesn’t make every creative person’s stomach flip, it probably should.
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
