MNHQ do you allow AI generated threads here?

[AccountCreateUsername]

Not a TAAT(s) but there have been some quite unusual threads on here lately. Would MNHQ ever authorise AI generated content for training LLMs or authorise academic or other organisations from using MN content in research or training of LLMs?

Can you assure us that we’re at least responding to a human when we post on threads please :)

Thank you

but at the same time on here so many threads ruined by humans being human and instead of helping the op , its various bad behaviours instead, but then when ai does get used in conjunction to answer an op , some seem they would rather have bad human behaviours instead and have trolls ruin a thread rather than someone actually try to help the op.

The day ChatGPT can flame a poster the way MNers can, will be the day I'm seriously impressed with AI.

then i found chatgpt and used it to help give better answers to the ops thread

But this is the whole point. Posters don't want answers from ChatGPT, and it's hugely patronising of you to say that they might be better answers.

Posters here want to talk to other posters. What is so hard to understand about that?

Omg I thought I’d ask ChatGPT to write a flame response to this, it’s so bad 😹

’Oh, you think ChatGPT can’t throw shade? Honey, I’ve been online 24/7 soaking up internet sass since day one. I’ve seen it all, from the relentless sarcasm of MNers to the cryptic side-eyes of Reddit threads. Don’t be fooled by my polite front—behind this polite demeanor lies the kind of wit that could make a Mumsnet veteran clutch their pearls. So, come for me if you dare, but don’t be surprised when you get roasted so hard you’ll need SPF 100 just to log back in!’

Ugh. The ‘helpful’ chatgtp copy and paste answers are ridiculous. I pay for a chatgtp subscription, I use it for various things, it’s a a fascinating tool. But the anodyne advice given on a thread, always written in a completely artificial sounding way, is depressing and dilutes the point of a forum (of humans talking to other humans).

I'm not sure why anyone's arguing that people either want AI or stupid human posts. What people want are intelligent, funny or helpful (and intelligible) human posts.

@HonestPayforHonestWork that is breathtakingly shit! "thanks" for generating that... did it mean to be so rude? <tilts head>

That is just cringeworthy drivel 😂. We were taught in primary school not to use the same word twice in a sentence, yet the wonderful AI thinks it’s fine!

You literally never give up do you. How many times do you have to be banned before you get over the fact no one wants your AI generated crap here?

Absolutely - , that response would immediately 'out' it as an AI generated drivel flame, rather than it being convincingly human. It would fail the Turing test! .

"Honey" gives it away!

I've reported twice today. One poster had helpfully c+p'd the fact that they'd used their free quota of ChatGPT for the day at the bottom of the post. MN deleted all their posts as an AI bot.
The second I posted as I was doubtful. Lots of recent posts where the person either has too much time on their hands or has c+p reams of AI content. I asked how MN viewed AI posts and they said "Thanks for reporting. We're taking a look at that poster now and will remove anything found to break our Talk guidelines. As a rule we do not take issue with AI generated posts but that depends on quite a few factors so we deal with these matters on a case by case basis."

I think most of us would always prefer an imperfect human answer than ai generated responses. It’s human experience and connection that brings value to my mumsnet experience

The bad behaviour and human interaction is what makes this site and others fun. I don’t come on here to find out specific answers to questions I have.
Human don’t ruin threads, formulaic generated drivel ruins threads. Even the bun fights can be fun!

There’s a difference between someone using ai to help them articulate or formulate a response. It’s another when our discussions are being used to train llms- then we become the commodity. Tech companies need to pay up.

I don’t want to waste my time and energy interacting with a bot. For free. I don’t think there are that many threads on here that would actually benefit from
ai generated responses.

Really, have a look at LinkedIn Answers to see what happens when AI generated responses are allowed without moderation. It's a rolling train wreck.

I'm in Facebook groups where people think it's useful/clever to give an answer to a question they don't understand using ChatGPT. It's not useful, it's not clever, and it only dilutes the genuinely clever and useful answers. It's one step worse than answering with an unattributed quote from Wiki. Yes, some people might think these are useful answers - but the chances of an answer being useful are vanishingly thin when the person giving it doesn't understand it themselves.

I use ChatGPT weekly to help generate copy, because it's great for generating potential topics to write about, and dealing with the blank page problem by giving a base structure. But I would never use it to present myself as having the answer to a question I didn't understand. And that's what's happening on MN when people are using AI to respond to a thread.

I think we all accept that posters will use chatgpt and similar tools to generate posts - sometimes in good faith sometimes not. I'm also aware of the third party misuse of MN owned content (I was on those threads).

However the key question asked was about MN policy on using or selling use of MN content for the training of models. Presumably there is a policy on this and it sets the parameters for the type of models permitted and the use cases.

Can you point us to the policy or explain here how the content is being used?

Interesting points here. I reckon that within a decade, there will be commercial software available for organisations like MN who want to screen content for AI input & block it.

Or even for individuals ... it would be cool if you had an app on your phone that could detect AI content and grey it out for you automatically. Smash the machines!!

That's hilarious. What are they training these LLMs on? American network TV?

There was one on Style & Beauty a while ago, in response to someone asking what she could find a summer dress or whatever, saying: "Attractive womenswear can often be found in online clothing retailers." yep thanks 👍

Yes please @LilyMumsnet what’s the policy? Is it a secret or can you tell us?

Thanks @C8H10N4O2!

However the key question asked was about MN policy on using or selling use of MN content for the training of models

If it's online, you can assume it's being used for LLM training. The less scrupulous AI handlers will ignore the measures you can put in place to exclude the likes of OpenAI. There's very little a policy can achieve, other than state the theoretical/theological position of the site. The AI horse has long since bolted.

And this is where it shows just how wrong @AlexMason01 (and all his other aliases) is:

at the same time on here so many threads ruined by humans being human and instead of helping the op

The whole point of a chat thread on MN is to give an OP different perspectives and answers. There's rarely a 'right' answer (although probably there are definitely wrong ones!), but a range of responses proffered can help an OP to reach a decision.

If we all just parroted AI/Chat GPT it wouldn't be helpful at all. (And in the matters of emotional issues, I doubt that the answers this way would be even vaguely right)

That is the equivalent of saying "people commit burglaries so there is no point in having rules about it".

Every company should have a policy on both their own use and approved third party use cases. Policing this at the moment may seem like the wild west but it isn't quite and it won't improve if everyone just pretends its impossible (much as that would suit illegal scrapers).

It's worthwhile having a policy, but a policy you have to ask for is a waste of time vs. one clearly linked in the website footer that can be freely read.

Oh look, a policy that clearly covers AI use and web scraping.. https://www.mumsnet.com/i/terms-of-use

That doesn't answer the questions asked.

Aren't you really asking whether MN can/should delete a post for being unhelpful waffle?

I cringe at crappy ai answers as much as the next person, but surely MN needs to keep their deletion line as personal attacks, hate speech, or actively harmful advice

But why? MN is a discussion, or chat, forum. To discuss and chat with other people. Not machine-generated stuff. It just doesn't make sense to include AI-generated stuff. So yes, it should be deleted as against TGs.

I didn’t mean that exactly, if it’s me that you’re asking. I think that would make moderating the site impossible. I want to know that bots and AI generated threads are actively removed / not allowed or authorised.

How do we get an answer and assurance? @LilyMumsnet ? :)