AI is great and everything, but I come here for human discussion.
I really don’t enjoy having to trudge through long, over-detailed responses that are clearly formatted as if they were cut and pasted from ChatGPT or equivalent.
They are too dry and too pretentious.
If I want an AI answer I can go get one myself easily. That is not why I visit this forum!
If you have a point to make, can’t you make it in your own words, or at least briefly?
Why are you subjecting everyone to have to waste their time reading such long indulgent posts?
I get that some people are dealing with a language barrier and it must sometimes be annoying that most of the business world uses English, but that’s no reason to waste people’s time with extensive posts lazily cut & paste. I’d much prefer seeing language mistakes as that often adds character.
It makes the forum less human, less engaging and really annoying.
If it was my forum I would moderate it heavily but it is not.
If you are translating using AI, can you please consider your audience and limit the length? Consider a prompt that makes the output more like you!
Opinion on heavily formatted, long posts that look as though they come from ChatGPT
Really annoying
Mildly annoying
I don’t mind either way.
I like them
I love them
0voters
(btw, the mandatory tag requirement is very silly in this Category and given the limited choice)
For this forum to be supportive for everyone these post are allowed. No one is forced to read them.
One should consider that many on this forum may not use English as their first language, be eloquent in their writing, be as expressive and thorough as an AI response, etc.
There couldn’t be a better tag than “ChatGPT” for this topic.
Occasionally, we come across posts rewritten by AI, and it can take some effort to distinguish between purely AI-generated content and AI-enhanced content. In my opinion, the latter category is perfectly fine.
Sure, as I clearly stated so you don’t need to repeat me.
But translation or tidying up grammar is not what I’m talking about.
I’m talking about cutting and pasting large, unnecessarily long posts that might even suggest that “the ideas” presented aren’t even those of the Poster, but of the LLM.
I might add I’m not criticising the moderators here in any way. This is a call to the community to act responsibly with these tools and to point out how such posts might not be received so well.
I think those are annoying when the author of the post did not think it through himself and just let AI answer so the quality of the answer is low. But when the reply is structured well thought through I don’t mind if it’s AI who wrote it.
It is easy to dismiss someone that has taken extensive time to answer with a level of high literacy, perhaps the reader carrying along with them the predisposition that a trivial question to them does not deserve such an elaborate answer, so clearly the thoughtfulness and accuracy in producing such an exhaustive depiction of the issues and arriving at a solution that is completely grounded in truth and reality must be mere seconds of pasting a question into a chatbot and receiving such a high quality answer.
I’m guessing that you’re referring to what are essentially generated reports of AI-enabled brainstorming sessions.
Technically speaking, I’d guess that it’s typically an attempt at refining an idea that the user couldn’t easily express in words, no matter the language.
I suspect that the fundamental issue is that the information density (or compression ratio?) of these posts is very low - because we (you, I’m guessing, and I) consider anything that the LLMs know to be non-information.
I’m thinking the conflict here is that some people either don’t know, mind, or agree with this.
If we could find a way to easily figure out what an LLM would find surprising, it should be fairly straight-forward to generate compressed summaries, maybe. Alt-text for text, or sorts.
I’m sure you’re aware, but a useful way to check for purely AI-generated content is to simply paste the original question into ChatGPT and compare its response to the forum post.
These types of posts also tend to be the ones that get removed.
One would think that smart answers were appropriately optimized to the best of the ability of the poster
Without removing too much of the context
A good post talks to many audiences because there are many readers
I wouldn’t consider myself alone to be that smart to not consider some AI collaboration in many contexts from research to evaluation to presentation based on the problem
Neither would I post anything first without at least thinking I understood exactly what I was posting first.
Use discourse features like details tags to hide out excessive content.etc
If you are going to use ai content have the ai return in that format so there is no additional human effort
Merefield is correct certainly in that there are some posts excessively AI
Maybe the discussion is best served by suggestions at best practice that could be referred to
Yes, the issue in part is due to the way an LLM will often generalise and elaborate and the result is a Post which doesn’t provide, imho, enough focus as a reasonable direct response to a prior participant, not paying sufficient reference to a prior participant.
For example, have you seen many “ChatGPT”-style posts quote individual sentences from prior posts and specifically address them?
Maybe behaviour like that can be prompted? But it would still be an artificial contribution.
I think your post touches upon an important aspect. We should not be using ChatGPT to undermine a forum within which many participants spend a lot of effort in drafting careful responses.
We should be showing respect to our fellow community member by spending the time necessary to formulate our thoughts and not use ChatGPT as a knee-jerk response or obtuse short-cut.
As stated, I think it’s totally positive to use AI tools to translate, refine language and grammar so long as the intention and ideas are preserved and not expanded upon artificially.
You also have to consider we are a world community, I myself right now I am collaborating with folks in England, Japan, Germany and other nations they all use it for translations.
Adding to general topic. To me it is like a person who uses dalle says “draw house” or some one who guides the machine to make art I see this as a medium of expression if guided .
How do you tell the difference though? We don’t know who anyone is. If they speak English or not and use a generation for the idea or to explain it. I see so many generations on OPs I ask questions and find out what they want or I page past and let someone better equipped than I deal with it.
