To think that multiple threads recently are being posted by computers/ AI so they can develop further understanding of human reactions?

[Frenchtoastie]

May be completely wrong, or this may be common knowledge?
Some threads posted recently are so bizarre and written in a weird way that I am sure it can’t be a human poster?
has anyone else noticed this or am I loosing my mind

No, lots of us have noticed and there have been several threads on the subject.

Well they won’t find any normal human interactions on here. Let’s hope they’re not trying to learn anything by posting on AIBU.

I'm sure a lot of them are, but unconvinced they are the bizarre OPs. Or maybe we just have very different ideas on what's bizarre!

The weirdest (IMO) threads are definitely not written by computers and I can't see researchers going to all the trouble of hiring writers who can post about peculiar situations in a consistently convincing style. Human beings just are weird, doing weird stuff all the time!

Those threads with totally inane OPs asking pointless questions, who don't come back to respond, though, I reckon must either be bored teenagers or bots.

Tell you what, though, I really miss the joke threads with increasingly bonkers OP posts. Some of them ran for weeks. Then everybody started saying they were bots or trolls, and they fizzled out 😢 Come to think of it, I should go for a nostalgic rummage in Classics.

I am so glad I am not imagining this

@EvangelicalAboutButteredToast 😂😂👍

Well it is 'amazing' all the people awake in the middle of the night with similar set out user names who post bizarre but very template like similar threads, I mean dont call me cynical or anything

No. Sorry to be snobbish but really, a computer would write better than most MN pps. 🤣

Exactly what I was thinking!

The alternative explanation is that they are Russian bot spies - who hopefully will be deterred from invading the UK due to the level of vitriol on AIBU.

Idea for a sci-fi film:

An AI that has been wholly trained on Mumsnet content is mistakenly placed in the body of a military robot and goes around executing all MILs, anyone who uses the word 'moist' and anyone who takes their dog for a walk in temperatures over 23 degrees.

It's been happening for about a year I think. I do think they should be highlighted as not real people by MN as you can see some people being very triggered and trying to help, which is a net negative impact.

It's also very boring when the badly written threads seem to all be hyper focused on one subject. I often get the idea that releasing a thread on here is someone's version of "getting the female perspective" so they don't have to actually survey women or pay them.

Our team at work lost 2 colleagues and senior management have told us we need to use ai and not replace them… but they’ve not explained how we should use ai, we have to figure it out (using the free version of copilot). If copilot was a colleague, they’d be on performance review right now! Execs don’t want to hear this and just think we need to learn it. So anything that improves ai is a good thing right now.

I liken it to being given nickle and glass and being told to make a diamond ring, then they are disappointed with the outcome and blame me.

copilot did make me an exercise plan that was quite good. Overall I’m so bored of broad claims of what it can do with no actual explanation as to how I make it do it.

Can you post an example or two, OP?

I thought this had been asked before and that MNHQ had confirmed they don't do it?

You sound like a crazy conspiracy theorist to me OP.

Much more likely is that some bored people are getting AI to write 'interesting' posts to get a reaction on here. I doubt MN would admit to it if it was them, but they might check that they read like they were written by a human. A bored random probably wouldn't care.

I do think there are a lot of fake posts, either weird trolls, AI and political AI. I don't think alot of sites do much about them because all traffic is good. 🙂 The quality has gone downhill though and turned it into a ridiculous divisive farce, which I don't want to be involved in. I think once people are wise to most online content being bollocks they won't bother with it eventually.

A lot of the disability bashing stuff is fake on here. They get removed because they are fake posts, rather than the content (disability bashing, particularly around disabled children doesn't seem to get sorted out on here unless it turns out to not be a genuine poster).

Although it's a bit annoying I am not going to complain to much or MNHQ will implement them pictures where you have to spot traffic lights / bikes / windows and such like before uou can post anything .! 🤔

I wouldn't be surprised after the OP I've just glanced at.

What really annoys me is when someone replies saying they'd run the OP through chat gpt or something and c&p the result as if it's anything useful. Don't outsource your brain ffs.

I'm afraid you sound pretty naive. I'm as far removed from conspiracy theorist as you can be. But , it has now been proven that there is huge amount of influence from Al and bots on many elections across the world. Brexit was interfered with, our elections, the US elections.
The rise in Reform is heavily AI supported.
To think they wouldn't be targeting MN is ridiculous.

Is this a case of any election that didn't have the result I wanted was nobbled by the Russians? I voted Remain and was a rejoiner for a bit until I got over it but iirc a major contributing factor of Brexit was two fingers up to that awful Cameron and all the rottenness he stood for.

Oh, did you mean the identikit political and social issue threads, @Frenchtoastie? In that case: definitely, though I understand they mostly come from Russian troll farms (they do use bots as well as humans).

They want to influence opinion, not research it.

I wish it wasn't the case, but it has been widely reported. As I m actually meant to be working but got distracted, I haen't got time to go through loads but here's one study (and there's loads more).

I agree about influence but I do think there's a lot of questions asked to enable training for better responses - if someone on here comes up with a good argument against a main manifesto point it might be tweaked or mocked in advance to rubbish it, for example.

Finding acceptable language for women seems to be where I can see them going with a lot of it; men don't instinctively know how to phrase things to fly with us (I'm going to call it pinkwashing in my head) so asking on here and having women call posters out as men highlights little tics and things that give them away. I think if they can post and not be detected they take it as a win and use those threads to nudge ideas and behaviours.

Interesting points. You're right about gendered communication - but they must have female trolls!

People keep saying this but it isn’t how AI works. There’s already plenty of text for scraping and training all over the internet including mumsnet, and that’s all that is needed, huge amounts of text. AI isn’t like some mad scientist robot doing experiments on people. AI might be being used by individual people in a spammy or trolly manner, but tbh lots of posts that people all clamour must be AI seem nothing like it to me. Do you use AI? I feel like if people want to spot it everywhere they should play with it a bit to get more familiar with what kind of text it does actually generate.