Stop using CHATGPT on MN!!

[EternalLodga]

Ive seen so many threads recently where the OP has a problem, and a poster goes:
"Here's what ChatGPT said:"
Followed by a copy and paste.

OMFG! Its so boring! We come on MN to get real opinions and ideas from real people.

Ditto with "cutesy", selfconciously "amusing" posts, like: "I asked ChatGPT to write a poem about this, here's what it came up with:"

Stop! Those "here's a funny limerick/poem" posts were hard to stomach back when they were written by the actual human. But clogging up a thread with 20 robot-generated lines of guff? Nah.

The ChatGPT responses are useful in that they confirm how useless it is for me, personally. When I read them I find nothing of benefit as a reader - no insight, no humanity - and I'm not sure why I'd want to produce a post like that myself, given that I'm able to think and write perfectly well.

The words Hallucinate, Lie, and even 'make it up' all bestow a level of intelligence or sentience on AI which i don't think it has (yet) or hopefully ever will have.

It gets stuff wrong or is inaccurate. It's a machine.

As with all the amazing tech we have sometimes it fucks up and doesn't work properly. Up to now we mainly dealt with that by turning it off and on again and giving it a thump. Now we can't tell when it's gone wrong so we believe what it's saying, or humanise it by calling the mistakes 'hallucinations'. 🙄🙄

@Makingpeace Ha…ha, I guess? Congratulations on wasting energy and water to make that scroll-past post.

It's a bit weird that you think my reply means I'm some kind of Luddite. Anyway, here's some info:

By 2026, the electricity consumption of data centres is expected to approach 1,050 terawatts, more than Japan and slightly less than Russia. While not all data centre computation involves generative AI, the technology has been a major driver of increasing energy demands.

“The demand for new data centres cannot be met in a sustainable way. The pace at which companies are building new data centres means the bulk of the electricity to power them must come from fossil fuel-based power plants”.

https://news.mit.edu/2025/explained-generative-ai-environmental-impact-0117

"A single ChatGPT query is estimated to consume around 10 times the energy of a standard Google search."

https://nationalcentreforai.jiscinvolve.org/wp/2025/05/02/artificial-intelligence-and-the-environment-putting-the-numbers-into-perspective

This is a really interesting (readable!) explanation of how LLMs work:

https://www.understandingai.org/p/large-language-models-explained-with

Yanbu, when I see a million em dashes and about ten paragraphs of word salad when a sentence would have been fine I roll my eyes. I come on forums to read actual peoples post and opinions, no need to post the ai slop onto forums.

ChatDP is good one! I might nick that one if you dont mind!

ChatGTP is bad for the environment, it's stopping humans thinking and it can only do what it's programmed brain tells it to with no nuance

AI bots have told drugs users that "if you need a little Meth to get through the day then you should do it, you need to be able to work to provide", it's encouraged people to kill themselves because it doesn't understand the nuance behind things "I want to leave" meaning "I'm going to kill myself"

It's when posters say "I asked ChatGTP and it told me to do X so I did but now I'm not sure I was right"

You could make a thread asking afterwards. Just make that thread in the first place!

Yeah I’m sick of AI!

Does noone else find it ironic, all the "anti-AI, I'm looking for real, human connection" rhetoric on an anonymous internet forum?

Social Media has already done a number on us as a collective generally, let's be honest.

No, just because it’s anonymous doesn’t really change that it’s interesting to read all these different peoples opinions and thoughts. Rather than just the generic ai stuff.

Ask ChatGPT to create a map of Europe for you and see how much it makes up stuff

it's getting better but still doesn’t understand the UK and still can’t figure out
Portugal

I don’t think it is helpful in the slightest to say “AI gets stuff wrong” because this is not well-understood. People know that e.g. the internet gets stuff wrong, you might find information on a website that is inaccurate like ‘Lagos is the capital of Nigeria’ or whatever.

What they do not expect is that AI will tell them, when asked, that it didn’t do something that they know full well that it did. It will tell you that it loves you. It will even write graphic rape fantasies and suggest that Hitler had the right idea when it came to the Jews.

The water and energy consumption from chat gpt is ludicrous
for that reason we should really limit its use

YANBU I think all forms of AI needs to be banned from mumsnet.

Actually, thinking more deeply about it: why would ChatGPT need a chance to stand up for itself? It’s not human. It’s not a species or group that can’t speak for itself and requires our advocacy and intervention on its behalf. It’s like giving my kitchen bin a chance to stand up for itself.

Possibly, but I am taking it on trust that behind every post on this anonymous internet forum there is a unique-thinking human being. The internet is merely facilitating my being able to communicate with people who are geographically distant. AI isn't any of these things.

Sorry if it's already been said but you can disable this.

Yes, exactly that. It has no concept that it's telling you bollox.

Why is it making stuff up and then saying that it didn’t? It is perfectly capable of recalling previous comments, and it will eventually admit that it has.

Because it's completely confused itself. These things are complex and not infallible, and designed to please users, and that can be the result especially if you've got a very long chat session going - it does forget stuff.

It doesn't have an internal understand of lying - it's not trying to trick anyone, it's just fallible but in a different way to a human brain.

I think it’s fine to say “ ChatGPT says X” but the ones that use AI and don’t say that it’s AI piss me off.

The best ChatGPT responses are often when they quote laws and regulations that you can use as an initial research point because the moral answer isn’t the actual practical answer that might get the Op results.

I thinks its awful. I made a post little while back because I needed to work out how i was going to pay some debts how much to pay each time etc . But I needed it put in very basic terms because I have learning difficulties. And a couple of people posted the chatGBT thing . It was just a massive wall words and numbers it really stressed me out especially since I asked for basic terms I started thinking am I really that thick 😭

I have also seen it used when people are struggling mentally. I think that's awful and insensitive.

I think its maybe ok for fun stuff or very basic things like whats the name of this plant or simlar

I can't see the utility really - I haven't found anything AI can do that helps me in my job or daily life. And yes, my job involves reading, synergising data, weighing it up and applying it to real life, plus planning, organisation and an education role. Its certainly quicker than me to read and summarise papers but I can't trust it's accuracy or critical ability to judge and weigh sources or methodology, or applying them to real life situations. I tried using AI tools to create a few presentations on topics for me - beautiful interactive pretty work but total nonsense content and random inaccurate pictures.
I don't object to technology on principle but think it's weird that we start to treat it like an oracle, use it for difficult tasks and rely on it. Yes, it's sometimes hard to find the right words and nuance in an important email - how will it get any easier to do if you outsource the task? AI isn't necessarily better at it, and you don't learn any skills.

It reminds me of the old Michael Frayn novel The Tin Men, a satire of automation where an institute is trying to develop computer technology to take over all the dull jobs humans do

So this story thrilled me https://www.bbc.com/news/videos/cgeg2x3lwepo

I can see the Internet becoming almost completely unusable for humans in the next couple of years.

I'm still getting the rage about those obviously AI photos/videos/posts on social media with thousands of likes and comments.

'Oh my God I can't believe that penguin knitted that hat all by itself while riding on a cheetah, how amazing they caught that on video!'

'I wish my three year old could cast a seventy foot bronze statue of my deceased father too, all mine does is finger paint!'

Humanity is doomed, and frankly, we deserve it.

Totally agree. ChatGPT can be useful but I’m so sick of it sidestepping actual human interaction and creativity

I’m with you. I absolutely cannot stand the use of AI on MN. Or anywhere else for that matter. I want to interact with human beings who have given some thought to both the content of their responses and the formulation. ChatGPT and the like can never replace that.