I have begun relying on AI and I don’t know how I feel about it.

It can tell you something stupid, you correct it and then it repeats the same nonsense again because it is only predicting the next word. It is not thinking about your problem at all.

I work in a pub, behind the bar. I can confirm that that is exactly how blokes in the pub interract 😄

https://www.nzherald.co.nz/nz/politics/act-mp-why-i-held-up-my-nude-photo-in-parliament/I4SI7ELBJZBTHP4SDHYMJDBWQ4/

but you surely asked that bot something you didn't know, so what exactly is it supposed to learn from what you didn't know?

Yes, but that is equivalent to asking a call centre operator to read your case notes. It is not comparable with a real human who might have had second thoughts, or gone away and read more on the subject etc. It is a facsimile of human interaction and people are easily fooled, it seems.

You can ask your ChatGPT to learn from previous conversations with you, this is something you can turn on in settings.

@chaosmaker Completely agree. And it's probably no accident that general xenophobia is on the rise as well.

There are loads of bollocks ai created nature photos on fb and probably on other sm sites that I don't use - cue thousands of comments under them with 'I didn't know that'. Of course you didn't, it's fake!

Edited to add this

Collective experience has been eroded ever since the advent of subscription tv for one. Then the internet made even more content available. When tv came out everyone had the same content and so a commonality of what they watched.
I think wfh is good but it is another way to estrange people from each other.

The bloke down the pub is not embedding his bullshit in all our daily lives and irrevocably tainting our collective archives and institutions.

Would you use the AI field mushroom guide taxguru? What reason is there for an unreliable field mushroom guide to exist?

Except that the bloke in the pub is a real person who lives in the real world and will remember the conversation days later so you can go back for clarification. With ChatGPT I get the impression that every interaction is a new thought process. There is no learning or understanding there. It can tell you something stupid, you correct it and then it repeats the same nonsense again because it is only predicting the next word. It is not thinking about your problem at all.

Oh FFS. Don't be disingenuous. Bloke in the pub versus AI is no kind of comparison, because bloke in the pub is probably five pints down, and his repertoire is highly unlikely to include legal citations that don't actually exist, but if it does your local barrister isn't likely to scribble it down on a fag packet and present it in legal proceedings.

Rightly or wrongly, things that look and sound credible on the Internet but that are blatantly wrong are being accepted because "Claude" says, or "ChatGPT" says, because AI is being pushed as more reliable than bloke in the pub with not enough caveats for those who have had their colours done, got a good recipe, or been offered meth by their virtual therapist.

AI is being pushed as a superior alternative to human interaction. Society is being further and further divided by the day. We're being subtly encouraged to believe we don't need other humans. At what point will AI agree with us, formulate a solution to the problem of our existence? I honestly think it could convince people to eliminate themselves in the long run, just by careful psychological manipulation. Maybe all those depopulation NWO "conspiracies" aren't that far fetched any more....

Just like the bloke in the pub then!

@Swirlythingy2025 if you think we're overstating the risks, would you use a genAI mushroom foraging guide? https://www.vox.com/24141648/ai-ebook-grift-mushroom-foraging-mycological-society

AI is more like: too busy to read a book? Just ask ChatGPT to condense it into a couple of sentences.

This is now a thing in academic publishing. No need to read any articles any more. What could possibly go wrong? 😒

Yes, I think Kindles are great tech – portable library! Not a fair comparison to AI. AI is more like: too busy to read a book? Just ask ChatGPT to condense it into a couple of sentences.

Everyone dies (Romeo & Juliet)
It was all a dream (Life of Pi)
Enemies to lovers, in costumes (P&P)

God, you could save so much time. Read thousands of books a day and still get your 12 hours in at the dystopian coalface.

Someone asked what I’ve used AI for this week and how it’s helped.

helped me draft my EHCP documents for my daughter. I had to send some info to a client ahead of a meeting - it helped me pull my jumbled up notes into a clear email that I was able to then adapt very quickly.

It helped me tidy up a document I had written for another client.

it helped me find a consultant and draft a letter for my elder daughter’s health needs.

all these things save me masses of time
snd make my day more productive and less stressful. I have an ND daughter who has PDA and Autism sometimes I’m completely stuck as to how to get something done - if she’s in the middle of a meltdown I can walk away and ask AI and it will give me suggestions - one of them always works.

everyday it can take hours for her to fall asleep. AI helps me with tips and stories. The list goes on. It’s made a huge impact to my mental health and physical - I’ve been able to do so much more in a busy day that needs doing reducing the pressure.

To be fair I love my Kindle. Pre Kindle I used to take 7-10 books away with me on holiday which took up a lot of room in my case! Now I can just take my Kindle.

And this is the problem with inventing a load of stuff we don't really need or actually want if we sat down and thought about it at all in any meaningful way. Give me a book over a kindle any day!

I would high five you in real life, if we were in a real pub , having a real debate because this - this that you have written with your actual tactile fingers, with the clear tone of bewildered exasperation, and relatable vernacular is absolutely how I feel. I would then buy you a real drink, we'd clink our real glasses, shake our real confuzzled heads and probably share our quiet existential crisis in companiable silence. I wouldn't even care if we were in a real beer garden with real wasps and we got caught in real drizzle, nor if real children were being noisy or someone's real dog ate my real crisps.

I want real, even if it's imperfect. That's my holy grail.

What is the fucking point of it all though? What is the pressing human problem we were all struggling with three years ago that is solved by ChatfuckingGPT? Fuck all, that's what. We are poisoning the well of human knowledge and society so that people like the poster just upthread can have their faces ground into the dust by their jobs a bit faster.

I hope you don't really think that those 2 examples have much in common.

Ai will make up sources. It will cite studies that were never carried out and experts who don't exist.

I agree that we humans should indeed approach AI with scepticism, but it's being pushed and introduced into the mainstream at break neck speed with those walking against it being accused of paranoia and being Luddites and being told if we don't get with the programme (ha) we'll be marginalised, jobless and on very short rations.

People's lives are busy - AI is being touted as a time saver. In that case the idea of going through it's output with a fine tooth comb kind of negates is usefulness.

i can understand your points, but thats a fault of humans if they take ai as gospel, plus with reguards to Ai thats only as good as the sources it uses so the websites that humans make need to be better content if the ai is using them etc
even humans rewrite history and yes should not does not mean it wont

I agree humans aren't always reliable, but the sources of information available to them should be as far as possible, not the product of some algorithmic fever dream programmed to "be helpful". Such as the ChatGPT therapist who advised a client struggling with addiction issues to have a little bump to alleviate stress.

As for perception and the re-writing of history, I take your point. But if we are psychologically being programmed to accept that AI output is a Holy Grail of some sort, we're in deep unchartered territory. If most learning resources are online, it's not just re-writing history, potentially it could be fabricating it entirely and claiming that living witnesses that things didn't happen are wrong. That's gas-lighting by technological means and the psychological impact is going to be awful.

Human memory as it is can be an unreliable and fragile thing, but some core truths and remembrance exist and should not be tinkered with for political or financial gain, or potentially just experimentally. We are not lab rats to be experimented on in this way without explicit consent.

Given that the High Court has recently had a conniption and put a stop to use of AI because it was making up citations, I still think it's too high a risk at this stage. Once a made up citation hits the Internet, repetition and reproduction makes the likelihood if verification far more difficult, especially if actual libraries and such resources are being made less accessible.