A curious reversal is taking shape as AI content floods the internet: work that is obviously made by humans is starting to command a premium of its own.

Writing in a new BBC analysis, Ros Atkins — presenter of the documentary series "The Value Crisis" — describes the phenomenon as a "human premium": extra value that audiences place on creative work precisely because a person made it. The evidence for it is showing up in music, writing, and the art market all at once. For more context on this story, see our ongoing artificial intelligence updates.

The Flood Nobody Can See

The scale of AI-generated content is no longer hypothetical. Music platform Deezer says 44% of tracks uploaded to its service are now AI-generated. Its research found that 97% of listeners could not tell the difference between AI-generated and human-made tracks. But the same research found that more than half of those surveyed said they do care whether music was made by a human — and want humans involved.

That gap between detection and preference is the crux of the issue. If audiences cannot identify AI content but say they value human authorship anyway, provenance becomes the product.

The Economics of the Human Touch

Daniel Susskind, an economics professor at King's College London, argues that creative work is fundamentally different from functional tasks when it comes to automation. He recalls telling a room of accountants that clients do not come for the "personal touch" — they come because they want their taxes done properly. Art does not work that way.

"If I walk into the Sistine Chapel and look at the ceiling, I think not only is that extraordinarily beautiful, but isn't it amazing that a human hand, flesh and blood like my own, came up with that," Susskind told the BBC. What audiences value, in his framing, is not just the destination but the knowledge that a human — one of us — got there.

The Kasparov Precedent

There is a historical template for what happens when machines surpass humans at a prized skill. When Garry Kasparov lost to IBM's Deep Blue in 1997, the fear was that interest in chess would collapse. The opposite happened: chess is more popular than ever, played by millions worldwide.

"It wasn't simply the moves on the board that people care about. It's the fact that it's a human being sitting behind them, moving those pieces along that matters," Susskind argues. Competitive chess thrived not despite computer superiority but alongside it — human contests became more, not less, compelling once machines removed any doubt about who calculates best.

Readers Want the Mess

The premium is visible in writing, too. Kyle Chayka, a staff writer at The New Yorker who covers technology, says AI writing is now widely considered "generic, obsequious and banal." As a result, he argues, "readers are coming to value very obviously personal writing, and maybe even writing that's messier, less structurally sound or composed or perfected, than they maybe once would've desired."

In other words, the flaws are becoming features. Visible human effort — idiosyncrasy, personality, even awkwardness — now functions as a signal of authenticity that generic machine polish cannot fake.

When the DJ Stops DJing

Atkins frames the question through a personal experience: a night in Falmouth watching a special guest set from Andy Cato of Groove Armada, artfully weaving tracks together — a performance that was, he notes, completely human. The DJ booth is where the question first got practical. In the 1990s, a DJ had to thread records together by ear with turntables and a mixer. Today, software syncs beats automatically, and both Apple Music and Spotify offer AI-powered tune-mixing options.

If you were out dancing, would it matter whether the DJ was really DJing? For live performance, the answer so far seems to be that audiences still pay for the person, not the playlist. But recorded music — where the product is a file — is where the flood has hit hardest, and where Deezer's numbers show how quickly the line between human and machine catalogs can blur.

Can the Premium Last?

The open question is durability. Human premiums depend on trust, and trust depends on verification. If audiences cannot reliably tell human-made work from machine-made work — Deezer's 97% figure suggests they largely cannot — then the premium attaches to whatever is credibly labeled as human, not necessarily to what actually is.

Disclosure Is Becoming the Battleground

The gap between detection and preference is pushing disclosure and provenance tools toward the center of the creative economy. Music services are experimenting with AI disclosure labels, and publishers face growing pressure to certify human authorship. Whether the human premium holds may depend less on taste than on infrastructure: the systems that let audiences know, with confidence, that on the other end of the work there was a hand — flesh and blood, like their own.

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →