The number of AI generated forum answers has increased for a while now. I am not referring to answers that employed AI to find answers and which have been carefully edited and fact checked. I am referring to lengthy and sloppy copy & paste AI generated answers.
Personally, I think they are quite annoying for a number of reasons. For example:
Quality They might sound convincing but more often than not they produce or include false and made up information. As a consequence, these answers become an echo chamber for misinformation.
Redundancy AI answers repeat information that have already been given in a thread
Lack of Originality Anybody could just ask AI themselves, there’s no additional value to it
TBH, these kind of AI answers keep me from following a thread and have a negative impact on my personal forum experience. So, here’s my question:
Is there a need to update the forum etiquette?
Should AI generated answers that have just been copied and pasted get banished from the forum? Or is this something we should just live with because it’s the new reality?
I would like to do something about this – but how would we go about it? There should be a way. Personally, I really dislike those verbose corporate-speak posts, and feel like one needs to fact check everything.
I think we need to be a bit more nuanced in the approach to this than trying to block all posts that appear to be AI generated…
I think having a conversation about it is a good idea. Let’s go
… waits for the first AI-generated reply to come in …
I would go further and say I’ve considered giving up the forums altogether, as I find myself wasting more and more time wading through the slop in the first place. Text AI-slop is relative easy to spot, but I recently spent a full 15 minutes watching a video that was so realistic, it took that amount of time for me to twig that the speaker hadn’t made a single mistake, or paused at all, and that the facial expressions and hand gestures, while occurring randomly, were identical each time. Worst of all, what initially appeared genuine and informative, was in fact AI-slop from persons with agendas…
Faced with such realistic stuff, we can hardly blame any forum user for falling for it, and posting a link without realising they’re spreading AI-slop, but perhaps we could have a mechanism to flag it as suspected AI-slop?
Then there’s the case where, through legitimate use of an ML-based tool, someone wants to post something and wants to openly label it as such … could we have a label or icon, or emoji for that?
I have zero interest in hanging out with bot generated content in this forum.
I can perfectly talk to the bots myself - I don’t need copy/paste services provided by other humans. I’m not that dumb (at least not yet ) and I’m oscillating between disappointed, disillusioned and angered, when I see such mindless copy/paste slop.
I’m here for the humans, that still use their own brains and hearts.
My thinking is, if we’re trying to help someone address an issue they’re encountering, and we Google something related to that to try and figure out possibilities, the first thing that comes up will be Google’s AI result. That might or might not be valid, but, at least in my case, I’ll read through it, and, if it seems correct based on my experience, I’m likely to paste it, but I’d probably also indicate that it came from Google’s AI and that it seems correct (or something to that effect to reflect that I agree). It would save me typing, and probably would result in a shorter response than if I did actually type one of my typically long-winded responses.
Mind you, I tend to look at questions/issues here as challenges, and, if they make me curious enough, I’ll try to figure out what the solution might be, even if I have no clue on the front end.
Yeah, I’m getting tired of the obvious copy and paste responses to posts. I am actually getting so tired of all this AI bullshit that I’m about to do bad things to bad people.
Discourse (the platform this forum is based on) provides customizable flags - this would be a viable option from a technical point of view. However, that also means more work for @steve and the other mods…
Of course, it is difficult to draw a line between reasonable use of AI research and mindless AI slop.
Maybe, the forum etiquette could define some basic standards like
AI generated passages
should be marked as such
should be kept to a minimum with regard to overall length, redundancies and info unrelated to the topic
should be fact checked before posted (to the best of one’s knowledge)
Maybe, this could help to save the forum at least from the worst kind of AI slop. Anything is better than nothing… This AI nonsense can really kill a thread: if you don’t read it you can’t keep up with the thread and if you read it you want to leave the thread.
I prefer your in-depth responses fold over something copy/pasted from AI or a web search.
And if you’re doing the answering to learn something yourself (which frequently is also my motivation), then at the bare minimum, re-phrasing the answer/concepts/reasons is a long known technique to make the learning stick better in your own memory.
So you’re not even doing yourself any favours when just copy/pasting.
@Scab_Pickens Haha, same here - I have to hold my horses, too…
About an updated forum etiquette:
I guess nobody’s eager to run around and mimic some sort of Anti-AI-Slop-Police. That’s not the point. Just knowing that there are some established ground rules might already do the trick at some point. It has helped to keep politics out of the forum, English as the official forum language, keeping conversations civilized etc.
I fully agree with this. It’s not so much that I hesitate to be a hardass and scold people from a perceived higher moral ground which in turn makes me a prick, but the disenchantment and disappointment in realizing that I’m not participating in a gathering of human beings and their own shared day to day experiences with subjects that we share a common interest for, but rather a solipsistic data-stream of self-proclaimed truth that does not leave room for truth outside of its boundaries.
I don’t know about you people, but I do happen to have experiences and come into knowledge of things that are not necessarily online. Knowing that I’m not special, I readily have faith that this holds true for most people. So the mere thought of having to prove if any human experience that diverges from the Geth Consensus is genuine or not discourages me and tires me gravely. (Around M.M.= 40)
I keep thinking “what’s the point?”, more and more often. I don’t need to post. Incoming transliterated proverb: Venice will not miss a sewing needle.
Having been recently asked if my post was AI because of grammar and spelling, I suppose I’m behooved to suggest that all AI is an accumulation of all the human input that trained the AI. Seems to me that the more important part is your ability to discern good data from bad data.
Does make me wonder though, how we’ve arrived at a reality where proper English is assumed to be AI.
I’ve posted questions in the forum before and when you are wanting an answer like “now,” a quick alternative is AI
In my experience it gives good answers in fact very good answers . You can sift through and take what you want. You don’t even have to say thanks but I do lol.
In fact you can have a good game of thank you tennis with the AI but it will always win.
But yes the copy and pasted bullet pointed AI replys don’t sit well with me ,they seem a bit patronizing even though the person who has sent them have made the effort where maybe just a short reply without the unesesesary bullet points would be better.
Probably because most people, especially we Americans (not sure if it’s the same in the UK), rarely use proper English? Too much phone use and small screens seems to make lack of proper grammar and spelling all the more egregious.
A few months back I saw a post somewhere on social media – I don’t recall if it was FB or IG – that was talking about telltale signs of someone’s having pasted a response from some AI (ChatGPT?). One specific thing it mentioned was use of the m-dash. Well, I use the m-dash all the time in my writing, even in writing that is just personal, and I really don’t use AI at all for writing. (My main use of AI relates to my use of Google’s search since, quite often, the AI response up top ends up providing the answer I was looking for, sometimes in the AI response itself and other times via the links referenced in that response.)
While I understand most of the arguments here, some nuances need to be made.
Many people here are not English-native speakers (as myself). AI is one way to help to write a text that has a coherent linguistic structure. Also, for many (even English-native speakers), the ability to synthesize an argument is limited, AI can help with this.
AI is a tool, if it gives me a way to express myself in a better and faster way, and gives the chance to help someone, why not? If I have more time, I prefer to write something by myself. But I don’t hesitate to use tools to help me with my writing.
With that said, there must be a way to prevent bot that would simply answer to everything on a USER-BASED forum. One goal of forums like this one is to create a sense of community, a way to share personal insights on different subjects. I read this forum almost everyday for this purpose.
) and I’m oscillating between disappointed, disillusioned and angered, when I see such mindless copy/paste slop.
.
fold over something copy/pasted from AI or a web search.