What to do when OP is using AI?

[samarrange]

On a currently active topic (which I won't link to here for now), the OP's original post and their every subsequent reply is very obviously written by AI. I put the text of the first post of the thread into an AI detector (zerogpt.com) and it says 100% AI, plus it's easy to can recognise the style after a while.

I have reported it and I'm interested to see what MNHQ do about it. The discussions are respectful, just weird. It's a topic asking for info, so most replies are from people who will probably just read the first post, reply, maybe look at the friendly (AI-generated) response, and not notice the pattern.

I guessed that it might be a non-native speaker, but then there are automated translation apps for that so that people can type what they want to actually say in their own language. As it is I don't know if OP is a human copy/pasting AI text, or a bot doing it all autonomously.

There is nothing that I can see in the Talk Guidelines about using your own voice. Using AI just seems disrespectful, but also, I am more concerned that there may be some "harvesting" ("scraping") of the site going on. (Again, I have reported that concern to MNHQ.)

I've reported a few posts (OP or not) that appear to be AI-generated - MNHQ take a look and if they think it's not a human they delete and afaik ban that 'user'.

Even if the post is innocuous it may be some sort of 'establish credential' attempt so I think MNHQ would like such posts reporting for them to assess.

I think that people come here to interact with other humans. If they are just typing at a bot that is here to harvest tips on a topic, for example to subsequently publish an article saying "The best 20 ways to cook spinach according to Mumsnetters" (the actual topic is not cooking, or spinach), I think that's abuse of their time.

Here is what the thread looks like (I generated the "OP" text with ChatGPT). There are about 6 replies from MNers and the responses from the bot/AI user always sound the same, give or take the precise nouns.

------
Hello everyone,
I'm looking to expand my culinary horizons and add more spinach to my diet. While I enjoy a classic sautéed spinach dish, I'm eager to try new and exciting ways to cook and enjoy this nutritious green. Do you have any creative or unique spinach recipes that you love? I'm open to all kinds of suggestions, whether they're for breakfast, lunch, dinner, or even snacks!
Thanks in advance for your suggestions!
Happy cooking,
OP

------
Hi OP, my favourite spinach recipe is Saag Aloo, an Indian dish with spinach and potatoes. It's vegan too. My kids love it.

------
Thank you for the recommendation!
Saag Aloo sounds fantastic! Do you have a favorite recipe or any tips for getting the flavors just right? I'm curious about the spice blend and any particular techniques for ensuring the potatoes cook perfectly.
It's great to hear that your kids love it too. I look forward to trying it out and hopefully impressing my family!
Thanks again,
OP

Good to know. I hesitated before reporting it because I didn't know if there was a policy and I don't want to waste MNHQ's time. (@mnhq It might be an idea to add "No AI" to the Talk Guidelines!)

I posted about something similar in Site Stuff earlier today - I wonder if we've been looking at the same threads? Turns out it was a real person, although I don't know why the style of their replies in those particular threads was so different to their usual style.

MNHQ responded "This is a long-time Mumsnetter and we don't believe they are being disingenuine at this point in time. We know AI can help some posters write out what they want to say."

Fair enough. Maybe the poster might like to know that she sounds like an overenthusiastic American, though. 😂

BTW @MNHQ doesn't exist. I mean, typing that will achieve nothing because nobody will see it.

You could report your OP and ask for the thread to be moved to Site Stuff.

If you suspect the Original Poster (OP) is using AI, verify the information for accuracy, and respond thoughtfully by adding personal insights or experiences.

Ask clarifying questions to ensure understanding and highlight any nuances AI might miss.

Encourage transparency by asking if AI was used, and promote ethical use while educating on AI limitations.

Adapt your communication to blend AI and human contributions, maintaining a critical and collaborative approach.

I don't think they have a policy with regard to ChatGpt.

I dunno - just lately it's been quite nice to read posts written be something intelligent for a change, even if it is artificial.

A lot of the humans are rather hard-of-thinking.

I don't think that is AI at all.

What you may find is people use chat GPT more and are adopting it's style.

You can always tell IMHO because it chucks in a word you just wouldn't use like 'endure' or something.

Ooo, I'd never thought about reporting them. Just found them deeply weird to read so would move on. Laughing at the examples - that's exactly the tone. Bland, overenthusiastic influencer type voice.

Again, the AI detection tool called it out as "100% likely AI".

But I suppose if we are now writing like AI, then AI has won. You won't be able to tell if anything online is a bot. Well done humanity.

Thank you for clarifying that for me. I appreciate your input and the information you provided. It's always helpful to have a better understanding of these things. 😉

Yes that's the thing. When I first came across I was like urgh and it jarrs so much. But I think the way the human brain often works is that we work to synthesise the old and new so it very quickly becomes something we don't notice or think is unimportant.

I did see MN are taking action to stop AI mining MN - imagine it having access to all the years of collective feminist wisdom that people have freely given on here over years, that just seems wrong. I mean I know we've freely posted but it was meant for people not machines or profits.

One issue issue is that, apart from the fact that most people come here to interact with other people - if I want to ask ChatGPT a question I can do it myself - some of these generated posts are coming from bots which might start off innocently enough to establish a 'posting history' but then be used to disseminate disinformation.

Another is that some of these AIs write utter rubbish but which can sound very authoritative. I don't know if it's been fixed yet but if you asked ChatGPT to provide references for a subject it would supply links... which were fake. AIs are not intelligent.