Declaring AI posts

[TeenToTwenties]

Would it be a good idea to add to MN posting rules

1. You may use AI to help formulate your opening post if you need help with organising your thoughts / want help with literacy but you must declare this at the start of your OP

2. A follow up poster may post an AI version of a long / difficult to comprehend opening post but must declare at the start this is what they have done

3. No other AI use is permitted when posting.

I think it’s blindingly obvious when someone has used AI to create a post. Especially when they are trying to be funny. It all sounds the same.

AI posts are blindingly obvious, but I think the reason for them matters. If someone is using AI because they have literacy or language issues or they’re copying & pasting information that’s different from a spammer.

Posters just need to be honest about it.

Because AI bots post on social media sites pretending to be humans, to farm interaction, generate content for other sites and generally manipulate people for some people's monetary gain. Humans using AI to write their posts are frankly wasting other people's time because they are contributing to the sloppification of the internet and reducing trust in the validity of basically anything.
I am a huge fan of AI and a big user for work and life purposes but using it to write a Mumsnet post is tragic.

ETA there is no point making rules about declaring AI use as the boys will just start doing that. Mumsnet needs to take a stand against it but they won't, they have inexplicably decided to embed AI in posts through the awful title generating option.

There's not much point in MN adding unenforceable rules. It is already harder to spot AI than many posters seem to think, and there's no way that MN mods would have the time and resources to check up on users' hunches.

It would have to be a matter of informal convention, with posters moaning about unhelpful AI screeds.

I think there are two ways in which AI can be unhelpful (in addition to the obviously abusive AI spamming, trolling etc). One is by making users' posts about their life issues too long and too polished, distancing us from the reality of what they are trying to say. The other is on less personal , more political/analytical threads, when people just throw in lots of long detailed content harvested from AI. Insofar as that sort of content has value, it is only in the actual conversation that the user is having with their AI. Outside of that context, it is just more dead text.

If it is not being used badly in those sorts of ways, it can be pretty neutral and unobjectionable. Better to call out the unhelpful nature of the post, rather than the manner in which it is generated.

Alternatively, if that sort of solution doesn't work, then I think the only other scenario is the slow death of MN. Moderation won't be able to rescue it from AI.

People want to read and respond to real posts from actual humans, having proper real-life discussions that aren't peppered with links, asking and giving advice and support, and whatever else.

Nobody wants to have a conversation with computer-generated drivel.

Absolutely.

Every single person who has the knowledge and ability to post on Mumsnet knows that GPT and other bots are out there. If they wanted to a bot to generate replies, they would ask one.

AI has no place on chat forums IMO.

AI normally tries to massage the questioner.

MN tends to lay the boot in.

So go to AI if you want agreement. Come here to feel like shit. If someone has been empathetic its probably AI?

I agree that filling a thread with AI-generated waffle can be irritating. But surely the issue is whether a post is relevant and contributes something useful to the discussion?

People can read newspaper articles and research elsewhere too, but nobody suggests we shouldn't bring those into a Mumsnet discussion.

Why should AI be treated differently? Surely we should be judging what is said rather than where it came from?

Filling a thread with AI-generated waffle is one thing; filling it with a multitude of links to other web pages/newspapers etc instead of answering the bloody question is something else entirely.

Bots pretending to be humans in order to manipulate discussions are a problem. However I use AI like yourself to formulate ideas, to prepare emails etc
You say you're a huge fan of AI and use it for work and everyday life. Presumably you don't regard everything you produce with its assistance as worthless or a waste of other people's time?

Isn't it more important, especially to the OP that any post is relevant, accurate and adds to the discussion, rather than simply whether AI helped write it?

There's a paradox that Mumsnet itself offers an AI-generated title option while members are arguing that AI has no place in posts.

In the example I was thinking of when I started this thread, the AI was definitely not in my opinion 'adding' to the discussion, it was overly long, telling us what we already knew and the 'opinion' was AI generate not by a human. Had the human read the AI they decided to generate, realised there was a specific insightful part, and posted just that part then fine, but they did not.

This is the big issue with AI, people cut and paste it without editing or nuance. This is a human discussion board, it needs humans.

Even if it were purely factual info being sought then posting AI answers still isn't really good enough if the underlying links haven't been checked, given AI's propensity to get the wrong end of the stick / hallucinate.

If OPs want AI answers they can ask AI themselves.

You’re a bit of a fun sponge aren’t you 😂

You cannot dictate how people use AI. MN can’t even dictate how people use AI when they post - it’s impossible.

There’s a better argument for your position than simply saying AI-generated contributions are too long or repetitive.

Mumsnet is basically a place where people come to exchange experiences, opinions and advice with other people. We don’t always want factual accuracy or logical quality of the answers. People want to know what another person actually thinks, based on their own experience, rather than receive a polished argument assembled with technological assistance.

That’s not unreasonable.

There's also a distinction between using AI to help express an opinion you already hold and asking AI to generate an opinion for you. In the first case, the technology is assisting communication. In the second, it may be substituting for the very human judgement that makes a discussion forum interesting. We all have opinions that AI might not know about us.

Even if the resulting argument is coherent and factually accurate, something potentially important may have been lost such as how the person feels.
Your point about volume is also significant. AI makes it extraordinarily easy to produce lengthy, plausible-sounding contributions with very little effort. I’ve been amazed at how quickly it can produce a play. That creates an imbalance between the effort required to generate a post and the effort required of everyone else to read, understand and respond to it.

A person can produce several hundred words in seconds, while another person may spend ten minutes carefully considering the argument. If that happens repeatedly, the discussion can become dominated by material that nobody has invested much thought in producing. Too much information can turn people off and people won’t read it.

That isn't necessarily fair to other participants.

And while humans can certainly be repetitive, inaccurate or verbose without AI, the technology changes the scale of the problem. Something that previously required time and thought can be done in a few seconds. Unfortunately speed and presentation can mean people trust the response without understanding it.

It's an interesting discussion about the use of AI to help you write replies to posts and compose thoughts on MN and people do just want to hear the human.

And yes. AI can produce long tedious replies

On the other hand, although blindly copying AI-generated material without checking it isn't helpful,you've said yourself that if someone reads the response, exercises judgement and selects something insightful, that's fine.

The key issue is not if AI was involved, but whether the poster has taken responsibility for what they've posted?

If a contribution is relevant, accurate and useful to the OP, why should it matter whether AI helped formulate it?

We don't dismiss arguments simply because someone used a newspaper, a website or a book to help develop them. I have been amazed at AIs ability to summarise radio episodes such as Cabin Pressure and it's provided insights to characters that I never considered before. Unfortunately AI can also make the most basic mistakes and ask things that it should really know.

And people can ask AI to produce short succint statements like this.

@cakeorwine I don't know whether you actually read what you just posted, but it was a longer-winded of one of your earlier posts, which you may not have actually read either.
Which all rather proves my point.

I've read a few posts lately that seem to be possibly AI. Usually long convoluted opening ones describing an unlikely situation. I just don't reply or read them.

Sorry but which earlier post are you referring to . Most of mine have been deleted

I actually like that idea tbh..I'd like it better as a filter so AI is automatically added. Not that there's such a thing for copy/paste.

They are spamming it seems.

Stop spamming

You've waited over 27 hours to tell me to stop posting?

19.14 & 19.52.

2 things

Both of them are replies to people's comments..Only one of them is to you so I am surprised you are commenting that I have repeated similar arguments to 2 posters. I am sure you know that on MN people often reply with similar arguments

Maybe you could find a way to compare the 2 posts. Look forward to your analysis

My thread, I consider all posts at least partly relevant.
Not going to waste time doing analysis of your posts when a good proportion are AI.

Well I have the right to reply to any post I want.

Did you find the long answer supportive of your belief?