Part 1 - Too Small to Sue, Too Small to License, Ingested all the Same
Part 2 - The Answer Without the Author
Part 3 - The Machinery of "No"
This is the first post in a series we're writing about original content in the age of AI, and it makes sense to begin with the shape of the problem as it currently stands.
When people talk about the AI copyright fight, the disputes that reach the news tend to involve organisations large enough to do one of two things: large enough to sue, or large enough to sign. The New York Times, Disney and Getty have chosen to litigate. Others have chosen to license: News Corp signed a deal with OpenAI reported at up to $250M over five years, and Reddit now licenses its corpus to Google for a reported $60M a year. What this tells us, more than anything, is that content does in fact have a price. At least, it does for those with the scale to negotiate one.
Too small to sue, too small to license
Below that line, though, sits almost everyone who actually produces original work: independent publishers, individual academics, NGOs, cultural archives. For the most part they are too small to sue and too small to license their work individually. And yet their content is ingested all the same.
The grievance isn't only about money
It would be a mistake, though, to assume the grievance here is only about money. When Taylor & Francis licensed academic content to Microsoft for a reported $10M (and Wiley booked more than $40M from its own AI licensing arrangements), the authors whose work was involved were neither consulted nor compensated, and had no way to opt out. The backlash from researchers that followed was less about the cheque than about the consent: about being used without permission, mis-cited, or stripped of the context that gave the work its meaning in the first place.
Credibility is the asset
And that distinction matters, because for researchers, NGOs and archives, credibility is the asset. The most acute risk of the AI era, for them, isn't lost advertising revenue. It's invisibility and mis-attribution: the quiet substitution of an AI answer for the visit to the source. A paraphrase with no name attached doesn't simply lose a click; it makes an entire body of work disappear in plain sight.
Machinery is emerging for the long tail
There is, encouragingly, some machinery beginning to emerge for what we might call the long tail: collective licensing initiatives such as the RSL Collective, TollBit and ProRata, alongside machine-readable rights signals that now carry genuine legal weight under the EU's opt-out regime. It is all still early, and none of it is yet proven at the scale where it reliably pays anyone, but the direction of travel is at least becoming clear.
First in a series
This is the first in a series of posts on attribution, consent and the economics of original content in the age of AI: what is actually settled, what isn't, and what small, mission-driven publishers can practically do about any of it.