AI research threads undermining MN credibility.

[CotswoldsQueen]

There is a long-running thread going on AIBU which - after several posters have called out the veracity of the situation being discussed - turns out was being done in the name of AI "research".

it is not clear whether this is an MN commissioned thing or if it was just the OP doing it.

Either way, it means that it undermines the whole premise of MN in that it is "by parents for parents". If we are getting to a point that it is 'by AI for parents" then the shared human knowledge, empathy & support that many posters come here for is no longer something that can be reliably found here.

So, @MNHQ is there anything that can be done about this? I used to be a paying member and dropped my subscription. I was planning on reactivating it but absolutely won't if I'm subscribing to a load of AI bots data mining.

I absolutely agree and was actually about to pose a similar question about the use of AI and the spread of misinformation on the site more generally. Misinformation is really dangerous - especially with health stuff, and some of the more contentious political issues debated on here. Personally I think there should be clear labelling on posts that have heavily used AI, and statements of fact should ideally have to be backed up by sources. I’d also appreciate clearer guidelines on how to challenge misinformation, is it something we report @MNHQ? Or challenge within the thread?

Out of interest, which thread is this - I'd like to take a look!

I've started seeing a lot more posts just lately which quote 'facts and figures' without any credible source. Eg: 'X percentage of people do Y' for instance.

The health boards horrify me. Full of people who know nothing but are happy to offer advice to OPs that could be actively dangerous in that situation.

@Westfacing it was one about a second wife wanting the first wife to accept that the second wife was having a baby. It's been deleted now the OP admitted the AI in a rather smug, "gotcha" kind of way and claimed it was for research.

Lots of people gave input from the first wife's POV obviously coming from experience and in some cases a lot of hurt and upset.

Yes, to the health advice but also in relationships, mental health etc. Where is MN's duty of care to vulnerable users?

It's deeply unpleasant and I would like MN to explain how this can be avoided, or even stopped, in future. It's in their interest as business to tackle the problem because the forum will lose its popularity if it isn't managed.

Hi there @CotswoldQueen . Thanks for getting in touch. We've just deleted the thread you refer to, as you have spotted - and no, it was nothing to do with us! We're not in the business of starting threads ourselves and definitely not with any AI-nefarious purpose.

We know that some of our users are a little suspicious of AI, understandably given we're being told it's going to turn on us all and eat us one day, and therefore are naturally suspicious when they see posts that they think might be written by AI. It's something that's come up a few times on Site Stuff threads.

Just to clarify, with threads like the one you mention where someone is clearly using AI to their own ends, we'll delete those. (We should add that we have absolutely no firm evidence that this was the case with that particular OP. They seemed to want to have their thread deleted quite desperately, which doesn't really track with their claim to be using it to train AI, and it's entirely possible this claim was simply made to encourage us to delete it. We'll never know for sure...)

However, with threads or posts where it's clear someone has used AI to help them draft a post but the poster is essentially 'genuine' (ie they just want a bit of help with putting a post together) we're ok with that. AI is being used widely in the workplace and for everyday tasks so we see no reason to disallow it.

As always, if you see anything that you're worried about just hit the report button and we're usually able to sort the wheat from the chaff. Or the robots from the real posters. We don't generally see much of the former type of post but we would always delete posts that did not seem to be here on a genuine ticket.

It should also be said, however, that much as some of us might mistrust, AI, we would never encourage anyone to make important decisions on subjects such as health or finance based on the say so of people on the internet and we do carry warnings about that on those sorts of forums. As we always say, we can never know with 100% certainty that anyone posting on the site is who they say they are so it's always best to exercise caution.

Thank you @CotswoldsQueen - I remember the thread. As you say many people did respond with advice and comments based on their own experience

It's best to stay away from family drama threads - thing is members often recognise the difficulty someone claims to be experiencing and feel the urge to be helpful

There's another one happening at the moment - starting from a very similar new-wife vs. ex-wife premise.

My issue is not people using it for grammar or more eloquent phrasing, its people like the OP of the thread I originally referred to actively using it for "research" or whatever data mining they are doing. How can you stop that happening?

Exactly that @Westfacing and that's what used to make MN special. Posters sharing their experiences from a place of kindness. This just was exploitative and, because it tapped into a topic that can be quite painful to discuss, cruel. I suspect that many of the ex-wives who commented on that thread had been treated in much the same way during their marriages or divorces, only for it to happen again on that thread when they were simply trying to help someone using the benefit of their own experience.

It's not fair to simply say stay away from the family drama threads. There are women on MN whose lives have been saved by the advice given from other women here when they were experiencing their own family dramas. For some people this place has been their only safe haven.

And as others have pointed out it is becoming more rife on health boards too which presents its own risks.

I can see that you delete them when they are flagged up, but how can MN be more proactive in this?

Can I just point out @CotswoldsQueen and @Raiderofthelostbattenburg that using @MNHQ doesn't actually work - it doesn't tag any of the mod team.