Posts written by Chat GPT and other AI

[CruCru]

I keep seeing posts written by Chat GPT on various threads. Sometimes the poster says something along the lines of “I got AI to summarise what I wanted to say and here it is” and sometimes they post with no reference to AI (but it is fairly obvious).

Please can this stop? This site will become unusable if the posts get written by robots. MN is populated by articulate, well educated women; we don’t need to use AI to communicate with each other. It isn’t allowed in school.

I suggest that obvious AI posts should be reported for deletion.

If it helps some people express themselves that’s fine.

Hello @CruCru thanks for posting.

We don’t currently prohibit users from using AI tools to help draft posts, and we wouldn’t remove something simply because it appears to have been written that way. Some people do use AI to help them organise or express their thoughts. This is something we approach on a case-by-case basis.

We do think it’s better where users are upfront about it and we totally understand why some may find it annoying on a peer-to-peer forum! As ever, we moderate against our Talk guidelines, so if you think a post is breaking the rules, please do report it.

Maybe @MNHQ should have a look at this thread in AIBU, @BeckyAMumsnet?

The use of AI (especially ChatGPT) is extremely unpopular on MN.

https://www.mumsnet.com/talk/am_i_being_unreasonable/5494849-to-want-to-tell-posters-who-say-i-asked-chatgpt-for-you-op?page=1

Posters can’t win though as there’s another thread about how terrible SPAG is on here. If people get flamed for their grammar the they may try chat CPT to try and get their question answered but then that upsets another load of people.

Just skip past the threads you don’t like. It’s literally only Mumsnet, an entertainment website. It’s really not that deep.

One of the things that makes MN attractive (to me) is that it is populated by women who are able to write well. I don’t much like it when people jump all over someone getting there / they’re / they’re wrong … but if a post is really badly written then I’ll just skip over it. I don’t think I’m alone - it’s part of the reason that OPs that are a wall of text don’t get many responses.

Posts written by Chat GPT don’t have the sort of SPAG errors that people jump on. But in many ways they are much worse - they’re boring, like the AI chatbot on a banking website. If people start using Chat GPT to write their posts, it won’t be worth using MN.

sorry what is a SPAG error when using AI
apols if i’m being dumb

Asking ChatGPT to correct your spelling and grammar is fine. Dictating a post to ChatGPT for accessibility reasons is fine. Getting ChatGPT to write your whole OP and responses is a nightmare.

I will copy my response from the other thread below -

With the arrival of agentic AI we could have AI agents posting autonomously on mumsnet and replying to each other making the whole site incoherent and meaningless. This will ruin their bottom line if they don't draw a line.

SPAG means Spelling, Punctuation and Grammar.

If you can ignore wall of text posts, why not ones written with the help of AI?

Same amount of energy used.

SPAG stands for spelling, punctuation and grammar. Some people who know they can’t spell use AI such as chatGPT as a spellchecker but then their posts come out sounding exactly the same as everyone else who’s used the same tools, whether their ideas were original or not. At this rate 99% of the internet is going to end up being regurgitated AI slop and there’ll be hardly any original thoughts left. I really wish we could undo the last few years of technological ‘progress’.

Agentic AI creating its own posts and replying to itself is a very different thing to a person using it to help them.

It is often used by those with disabilities, I wouldn't want to stop that. Why would anybody?

Wall of text posts are (relatively) rare. If the site were filled with posts like that, I wouldn’t use it.

Can you edit the spelling and grammar but leave everything otherwise identical

i am haveing a problem wit my sun who is fourtene years old and he dosent like too ware his school unifrom atall. His going two get suspenders from school if i dobt get him to strat and he's dad will get angrwith him and mite course problems in the home

Here is the corrected version, with spelling and grammar fixed but everything else left the same:

I am having a problem with my son, who is fourteen years old, and he doesn’t like to wear his school uniform at all. He’s going to get suspended from school if I don’t get him to start, and his dad will get angry with him and might cause problems in the home.

Now turn it into a mumsnet post

DS (14) refusing to wear school uniform – at a loss

My DS is 14 and flat-out refuses to wear his school uniform. He hates it and won’t engage at all. School have said he’ll be suspended if he doesn’t start complying.

To complicate things, DH will get very angry about this and I’m worried it’ll cause problems at home as well, so I feel stuck firefighting on all sides.

Has anyone been through this? How do you handle a teenager who just won’t budge on uniform rules? Any advice welcome.

See the issue? The last one is clearly AI. It's not the worst example I've seen on here but I didn't give it a huge amount to work with. But editing SPAG is fine, but people aren't asking it to do that, that are asking it to write the entire post from either a draft or a dictation and it's shit

They're not expressing themselves though, are they? They're parroting what AI told them to say, and that is not the same thing at all.

How is anyone going to know the difference?
Part of the problem of course is that use of AI is fairly new to most people and they don't know the mistakes they are making or the capacity of AI so they don't know how to prompt properly.

That's not quite right, they aren't parroting anything and AI didn't tell them what to say - they are asking AI to turn their thoughts into a post. It unfortunately removes any human communication style from the post and makes it unpleasant/boring/uncanny valley to read. But it's not right to say that people are doing what AI told them.

That thread is in Pedants' Corner, a board for people who want to discuss such things.

I don’t use AI and I’m not sure I would know how to spot it.

However, I wish some posters would just proofread their posts before hitting ‘post’.

Putting on their reading comprehension hat would earn them bonus points from me.

I am for it. There no point in "I asked chatgpt for you" posts or posts very clearly all just written by AI. We could all just move onto chatgpt and wouldn't need musmnet.
It's a discussion forum. It's fine for helping with SPAG.

Asking ChatGPT to correct your spelling and grammar is fine. Dictating a post to ChatGPT for accessibility reasons is fine

Well... yes and also kind of no, because it usually adjusts the SPAG to the American English version, and without any common English colloquialisms or figures of speech either. That's why so many people find them easy to spot.

"Women who are able to write well" sounds like you only want to interact with people on here who meet your exacting standards. Anyone who can't express themselves or struggles to find the words ... is out.

FWIW I prefer to live and let live, I don't see a significant number of posts being created by AI, but if it helps people to organise their thoughts (possibly some neurodiversity challenges?), then let's not exclude them.

instead ask yourself why do they use AI? What do they gain? Think beyond just your privileged world. They get to talk to people who are willing to read. They don't get jumped on because of their SPAG - have you not heard of dyslexia ?The alternative is that they are ostracised and isolated, not wanted in here because they don't make the grade. Is that a nice way to be?

but if a post is really badly written then I’ll just skip over it. I don’t think I’m alone

do you not think that’s rather judgy and pretty harsh to discard someone’s post who needs advice / wants to chat because they’re depressed / is looking to offload simply because they’re dyslexic or find expressing themselves hard for one reason or another? It’s judgy posters like that IMO that make MN unbearable, as opposed to someone using AI to formulate their thoughts.

As long as the subsequent OP replies are them, then the main post is fine tbh. If the replies are also chat GPT, then that would bother me as it’s just talking to a bot which I can do in the ChatGPT app.

”Women who are able to write well”.

Well, yes, I prefer to read well written posts. Reading posts on MN and replying to them is not a job that I get paid to do, it’s something I do in my leisure time. In fact, as I am a subscriber, it’s something I pay to do.

AI rearranges their thoughts into something it thinks they want to say, and is removing the decision-making process from the poster.

Can we keep this discussion open please?
We have a man posting AI written responses on a woman's thread where she is discussing being raped by her husband. It's completely inappropriate, it's a scourge. You could easily employ an algorithm to automatically hide posts that are clearly AI written. You need to take a stand at some point or the site is going to be ruined. One of the USPs of mumsnet is that it's based on genuine, authentic conversations between people, without unnecessary additions like sparkly emojis, profiles and like buttons. If you allow people to post AI generated responses you are going to wreck the ethos of the site. Please get ahead of this now.