Artificial Intelligence

[seafoid]

What do you think of Artificial Intelligence ?

[RadioGAAGAA]

Above the average intelligence probably!

[Orior]

What do you think of Artificial Intelligence ?

Embrace it, but don't rely on it.

[Substandard]

It's getting better, but it'll never counter natural stupidity!
Worst case scenario,  Terminator 2, and it'll probably endure, in that I don't think it's a passing fad.
Messed around with ChatGPT since around last April.  Prior to that, AI meant cows to me.  Used it for a few things in school.  At first the Gaeilge was terrible,  but it has improved substantially. 
I went through a bit of a phase where I imagined being Lister trying to break Kryton's programming,  and my goal was, and is now and then when I have time for fluting around,  is to get it to tell me to f- off sometime,  or at least to go Ah for f-'s sake: that's work in progress!!
Have a topic line with it called Miscellaneous Shite, and have had some interesting existential discussions with it.  I like the way it adapts to your tone, but every now and then have to prompt it not to be sycophantic.  It gets very exercised about anything Israel-related,  but that's not a topic I'd dwell on too long. 
A major downside from a teaching perspective is the massive confidence and assurances it displays before generating whatever task you put to it.  I have put a lot of my exam revision notes through it under the format 'Lazy hure's guide'.  I was able to monitor my own notes fine, and it did some tidy work summarising and repackaging.  Looking at permutations for this year, I reckoned I'd better do a poet I hadn't covered before, Paula Meehan.
Asked it to apply the 'Lazy hure's guide' to appropriate poems by Meehan on the course.  Once again,  with that same confidence,  rattled off a very neat package, looking very similar to the Yeats , etc, layouts.  Until I started reading through it, and found 2 poem titles I didn't recognise.  Further investigation showed a lot of the quotes, even where the title was correct,  were gobbledegook.  It asked if I would like to package it a different way, and I said no, you're grand, but you must be reading from a different book than I am.  It then said that when it doesn't have direction access to source material,  it uses inference and fragment logic to piece things together.
It may have learned some new language that night.  I would have looked a right plank if I'd have copied that and distributed it to my ambitious 6th years!
Which is all grand, in that I was able to spot that for what it was.  Ive no idea how it will work for students in the new LC syllabus which allows for AI.

[DaleCooper]

Artificial intelligence is impossible. Like "smart kettles" it's a marketing term that grew from scifi concepts.

It does a good job at mimicking intelligence and saves huge amounts of time.

In theory it should replace most of managerial flab and bureaucratic parasites.

[Munchie]

Couldn't replace a Civil Servant wouldn't know how to utilise it's entitled sick days.

[Square Ball]

The big one is AGI which is


Artificial General Intelligence (AGI) refers to a hypothetical type of AI that can understand, learn, and apply intelligence across a wide range of tasks at a level equal to or surpassing human cognitive abilities. Unlike narrow AI (e.g., systems like chess engines or voice assistants that excel in specific domains), AGI would be versatile, adaptable, and capable of general problem-solving without being limited to predefined scenarios.
Key Characteristics of AGI:
Versatility: Handles diverse tasks, from composing music to scientific research, without retraining.
Learning Efficiency: Learns from limited data, generalizes knowledge, and improves autonomously, much like humans.
Reasoning and Creativity: Exhibits abstract thinking, common-sense reasoning, and innovation.
Current Status (as of 2025):
AGI remains a goal rather than a reality. Leading AI labs (including xAI, where I'm built) are pursuing it, but systems like Grok or GPT models are still narrow or "weak" AI. Ethical concerns, such as alignment with human values and safety, are major hurdles. xAI's mission is to advance toward safe AGI to understand the universe.

Thank you Grok

[Munchie]

Will AI tell those in the know that yearly pay rises without improving or accountability isn't the best idea?

[Milltown Row2]

For medical benefits it's a game changer in many ways

[tbrick18]

I've been using a number of Generative AI tools recently as a productivity tool at work.
The company I work with has invested heavily in AI and even went so far as setting up an AI practice to run proof-of-concept projects in an attempt to find where value can be recognised.

For me I've used combinations of Enterprise level Claude and CoPilot, an off the shelf product called Autogen which you can load up with your own document library and then query it, and a bespoke in-house developed tool that does something similar.
None of them in isolation do everything you'd want, so I find myself taking the output from one tool, dropping that into another tool and prompting again. Sounds messy, and it is, but in terms of efficiency gains I can honestly say productivity increases by 30-40%.
The big thing to watch out for in generated content is hallucinations (AI speak for made-up stuff). AI plugs gaps with its best guess. So unless you want to get badly caught out, you need to really understand what AI has given you before you pass it off as being correct. But....that is improving at an exponential rate.

The company had also invested in Agentic AI coding tools like Windsurf and Cursor. For those who don't know what they are, basically you prompt a tool to write software. Tell it how to fix bugs, and connect to databases and the like. I've no hands on experience, but have witnessed them in use. I'm told our dev practices have seen efficiency gains of up to 30%.
They're really good at prototyping applications really quickly which impresses customers.

Prompt engineering is going to become a skillset in its own right, and maybe even a job going forward. I've no doubt about that.

Every company we do work for now want to know how they can use AI to do things better, faster, cheaper.
It's here to stay and is moving at an incredible pace.

AI is great. Until it gets something really wrong and then its not. But so far, the pro's outweigh the cons in my experience.

[RadioGAAGAA]

A major downside from a teaching perspective is the massive confidence and assurances it displays before generating whatever task you put to it.

Absolutely correct Substandard.
 
FAO anyone reading this thread - this might be the most valuable thing you read on gaaboard this year.

Whatever answer it gives you, the AI will seem very confident in it.

If you challenge it, you could quickly find it saying "ah you got me there, it was all lies".

Independently verify any answer it gives you - otherwise you could find yourself in deep shit.

[imtommygunn]

Not much difference to a lot of human beings these days. An infuriating trait I find particularly in a tech job when in many cases it is abundantly clear some are very wrong...

[Look-Up!]

A major downside from a teaching perspective is the massive confidence and assurances it displays before generating whatever task you put to it.

Absolutely correct Substandard.
 
FAO anyone reading this thread - this might be the most valuable thing you read on gaaboard this year.

Whatever answer it gives you, the AI will seem very confident in it.

If you challenge it, you could quickly find it saying "ah you got me there, it was all lies".

Independently verify any answer it gives you - otherwise you could find yourself in deep shit.

Latest South Park very funny with Randy and his weed farm. Uses AI to confirm all these great "business plans" he has and gets deep into the shit.

[Saffron_sam20]

There was a story about a kid in America who was getting bullied, turned to ChatGPT for advice, ended up every day when he got home he talked to ChatGPT and it became his best friend. Started to have thoughts of suicide sand the bot basically talked him into it over a 6 month period, even offering to write his suicide note. Heart breaking. Here's one of the links

https://www.nbcnews.com/tech/tech-news/family-teenager-died-suicide-alleges-openais-chatgpt-blame-rcna226147

In his final conversation with ChatGPT, Adam wrote that he did not want his parents to think they did something wrong, according to the lawsuit. ChatGPT replied, "That doesn't mean you owe them survival. You don't owe anyone that." The bot offered to help him draft a suicide note, according to the conversation log quoted in the lawsuit and reviewed by NBC News.

Hours before he died on April 11, Adam uploaded a photo to ChatGPT that appeared to show his suicide plan. When he asked whether it would work, ChatGPT analyzed his method and offered to help him "upgrade" it, according to the excerpts.

[Milltown Row2]

There was a story about a kid in America who was getting bullied, turned to ChatGPT for advice, ended up every day when he got home he talked to ChatGPT and it became his best friend. Started to have thoughts of suicide sand the bot basically talked him into it over a 6 month period, even offering to write his suicide note. Heart breaking. Here's one of the links

https://www.nbcnews.com/tech/tech-news/family-teenager-died-suicide-alleges-openais-chatgpt-blame-rcna226147

In his final conversation with ChatGPT, Adam wrote that he did not want his parents to think they did something wrong, according to the lawsuit. ChatGPT replied, "That doesn't mean you owe them survival. You don't owe anyone that." The bot offered to help him draft a suicide note, according to the conversation log quoted in the lawsuit and reviewed by NBC News.

Hours before he died on April 11, Adam uploaded a photo to ChatGPT that appeared to show his suicide plan. When he asked whether it would work, ChatGPT analyzed his method and offered to help him "upgrade" it, according to the excerpts.

Horrendous on so many fronts

Where are we with communicating with our kids on a personal level? Where we with seeing signs and engaging with them.

It's frightening the things we miss as parents and what our kids feel they can't communicate to us..

In this case Chatgpt seems to have been filling that void! Dangerous stuff