Exactly, the main issue isn't the LLM creating or confirming the report. Rather, the maintainer has no idea how it was prompted, and the results may be completely wrong. Some LLMs also have a bad habit of trying to please the user, confirming their biases.