If you look through the history of the Show HN category, you will see a near-endless stream of "analyzing HN discussions" prior arts that may offer some assurances. I'm not suggesting removing usernames because anonymity is involved — it is not! — but, instead, to specifically focus viewers on the substance of the comments rather than the who of them. That's also why I chose something from two years ago, so that there's no appropriate reason to witchhunt over the past. (Some may still, but nothing short of evaluating AI-generated data will stop them, and you can't make a reasonable case using AI to evaluate AI.)