I'm pretty sure IF a copyright lawsuit went sideways you would still be open to litigation risk, just hiding the evidence.

What you're doing would fundamentally be similar to copyright theft, using 'someone' else's code without attributing them (it?) to avoid repercussions

Obviously the morals and ethics of not attributing an LLM vs an actual human vary. I am not trying to simp for the machines here.