Who owns AI-generated content?

Ownership feels simple until AI gets involved. If you write something yourself, it’s yours. That’s the assumption most of us start with. But the moment a system is generating the output, that clarity starts to slip.

Because the result isn’t coming from one place. It’s shaped by a prompt, built on training data, produced by a model. There isn’t a single point you can easily point to and say, that’s the author.

So we start asking slightly awkward questions. Is it the person who wrote the prompt? The company that built the model? The people whose work sits somewhere in the training data?

Most of the time, none of that really comes up. People just use the output and move on. It feels like theirs, so they treat it that way. But that assumption starts to wobble the moment it matters, when money’s involved, or attribution, or originality.

And the more this kind of content gets produced, the harder it is to pretend the old idea of ownership still fits neatly around it. Maybe the question isn’t just who owns it. It’s whether ownership still works the same way when creation isn’t coming from a single place anymore.