LogFAQs > #969967511

https://www.theverge.com/2022/12/5/23493932/chatgpt-ai-generated-answers-temporarily-banned-stack-overflow-llms-dangers

For stack overflow it's particularly about generated answers that may look real from a cursory glance, but fall apart in debugging (ChatGPT can generate stuff in any coding language, but it just knows what common codes look like, not actual logic). This makes it really hard to review answers that are being flooded in since it takes a lot of time to confirm the bullshit doesn't work.

But ive seen it now for advice style forums too. ChatGPT generations to questions in paragraph form are very similar to the websites you see when you google a tech problem. Lots of restating the problem + wordy but empty platitudes, but of course there's no substance or insight because the AI is just generating what a safe, unthreatening answer should *look like*, not if it is helpful.

It even happened to me on 4chan's /adv/...