The 'bug' here is whatever post-processing step or system prompt is in place to steer the model away from doing this.
One could argue nobody should be allowed to claim it. It just exists.
Even if the person can't claim the copyright of the image produced they ARE responsible for the use of their tools and what they do with the output.
In this case, they released an image with someone else's signature on it. That is wrong, the person should take the blame for that.
The person releasing the image may take it up with the AI service that their tooling led them into making such a mistake. But good luck with that in court...
Somebody “made something.” But just because you do something doesn’t mean you get to claim whole ownership of it and get to sign it with your name. Plenty of examples in life.
LLM generated picture with artists signature. That's no problem.
Person takes this picture and proceeds to share it around is the problem.
What annoys me is when "LLM hacks website!". Uh, it just went and did that completely unprompted? Nope.
Somebody ran it on a computer somewhere, gave the LLM some kind of prompt+access combination - however precise or vague - and THEN the LLM did a thing.
Power tools are well known for making human bits go missing. But nobody says "The angle grinder did it on it's own! Nobody is at fault!"
Maybe the person wasn't trained. Maybe it's the wrong tool for the job. Maybe the manufacturer of the tool is at fault for selling a product without appropriate safety guards. But the LLMs aren't doing self sustaining thinking for themselves just yet.
If the machine is like you, the machine is a forger. The machine is not like you, it is simply blending the work of others to order. Adding someone else's signature is simply part of that statistical process.