Absolutely, evidence of abuse is secondary to the actual abuse.

Plus, the fact you could use/make AI/LLM/etc generate nefarious content that is hard to tell is fake, tells you the abuse isn't even what they are interested in.