There’s a bit more nuance to your example. The company is liable for building a tool that allows plagiarism to happen. That’s not down to how people are using it, that’s just what the tool does.
So a company that makes lock picking tools is liable for when a burglar uses them to steal? Or a car manufacturer is liable when some uses their car to kill? How about knives, guns, tools, chemicals, restraints, belts, rope, and I could go on and nearly use every single word in the English language yet none of those manufacturers can be sued for someone misusing their products. They’d have to show intent of maliciousness which I just don’t see is possible in the context they’re seeking.
The reason GPT is different from those examples (not all of them but I’m not going into that), is that the malicious action is on the part of the user. With GPT, it gives you an output that it has plagiarised. The user can take that output and then submit it as their own which is further plagiarism but that doesn’t absolve GPT. The problem is that GPT doesn’t cite its own sources which would be very helpful in understanding the information it’s getting and with fact-checking it.
While GPT was trained on the material it does not produce plagiarizing results. It can have reused phrases but only because those phrases are reused across multiple examples and not from a specific work. It learns like b comes after a, c comes after b, d comes after c and then will sometimes reproduce ABCD because it’s normal for that to be used within the context. It is not plagiarism but more akin to the human capability of guiltless probability. If it’s plagiarizing then it’s doing so by coincidence due to context.