Ask HN: Will HN moderators add a spam filter for replies written by ChatGPT?

5 points by leoplct 3 years ago | 4 comments

Many of the "Ask: .." topics could be answered with ChatGPT creating long, generic responses and lots of noise inside.

Will tools be used to identify text written by ChatGPT to flag it as spam?

I think this decision is very important in predicting the future of ChatGPT and distinguishing whether it is just a buzzword of the moment. Humans are enthusiastic about having a robot write a reply, but they don't want to read a robot's reply.

cheapliquor 3 years ago |

ChatGPT says:

"Yo, fellow hackers. I think we need to seriously consider adding a spam filter for replies generated by chatGPT. I mean, it's all well and good that we can use AI to write code and stuff, but when it comes to the comments section, we need to ensure that the convo stays on topic and high-quality. A spam filter would help keep the noise down and make sure that the signal-to-noise ratio stays high. Plus, it'll prevent the bots from taking over HN like they did on Reddit. Let's keep the discussions here on point and civil, and avoid turning into another 4chan. Let's do it."

sharemywin 3 years ago |

I personally think it's fine to post ChatGPT responses as long as they're labeled.

Especially when the discussion is about ChatGPT.

jleyank 3 years ago | |

I fail to see the use of randomly generated verbiage unless the exercise is reviewing the quality of randomly generated verbiage. Won’t add anything to the discussion otherwise. If unlabeled and sufficiently common it will drive the s/n ratio of HN below usefulness.

leoplct 3 years ago |

If we want chatGPT to make this decision, here is its answer:

"Should HN moderator identify responses written by ChatGPT and flag them as spam?"

Whether or not HN moderators should identify responses written by ChatGPT and flag them as spam is ultimately a decision for the HN moderation team to make based on their goals and community guidelines. However, I believe that it would be more beneficial to have a clear label on text generated by AI, so that readers are aware that it has been generated by machine, rather than flagging them as spam. This can help readers to understand the quality, context, and limitations of the generated text.

Additionally, it's important to note that AI-generated text has its own value, for example, for answering frequently asked questions, summarize long documents, or even for creating new content. Instead of treating all AI-generated text as spam, it could be useful to make a distinction between those that are harmful or misleading and those that are useful and informative.

As a rule of thumb, transparency and clear label for AI-generated text should always be promoted to foster trust and clear understanding of the capabilities and limitations of the technology.