Unconscious

 

Like a lot of disgruntled folk, I'm dealing with the absolute la-la land bullshit of a company falling head over heels in love with AI. More accurately, falling in love with a suspiciously underperforming mathematical box of tricks that performs no neater a trick than just combing through a bunch of other people's hard work to pretend it's all its own. 

Now, I can fully understand why the people who are in love with this thing (or those things, as just like the damned hydra you cut off one of these things and another 9 spring up in its place) continue to try and push them, and also continue to throw money at them. They are falling for the bullshit, and falling for some slick-assed (and usually American) sales pitches from folk who are experts at selling The Emperor's New Clothes, Snake Oil, or whatever you want to call vapourware these days. 

In some sectors, particularly in my line of work, the pushback is real but the pushback is often squashed under a blanket of policy that basically says "Use this stuff, learn to love this stuff, or it's your job". 

I recently talked to my Nephew, all of 10 years old, about learning. He's a boy, he's 10, and his attitude to school is "I have to be there, that's the only reason I am there". But this boy is a knowledge sponge. He loves to hear stories, he loves to learn, he absolutely SOAKS that shit up as quickly as you can talk about a subject, and he'll come back with 10, 20, 30 more questions about that subject to keep you on your toes. We were walking around a museum and he became obsessed with the amount of gold on show, even hosting his own little heist movie in his head about the things he would break in and steal from the museum, melting them down into one huge glorious block of gold (to do what with? Who knows what goes on in the head of a 10 year old). 

I loved that idea, but I digress from the original purpose of this blog post - to lay into AI and put the boot into its insidiousness. 

I hate seeing folk try to push an agenda that it will somehow solve the world's ills. It is the chief cause of many of them (the amount of water data centres servicing AI use, the amount they are contributing to global warming has been measured and is palpable - all for the sake of someone shoving a mindless prompt into an image generation algorithm for a brief 10 second giggle or two). More serious is the inaccuracy of the thing. Our team (with a collective experience in Educational IT Support that spans a couple of centuries if you add it all up) is constantly questioned on the answers we give "Because Copilot told me that you're wrong, you are wrong". That to me is the most insulting as all of us have gone through professional training, and have attained real-world experience that a crappy algorithm would never manage, instead acting like a cuckoo in the nest by scraping and stealing other people's information, accurate or otherwise. 

Companies are beginning to realise that AI deployment is expensive, and that the returns are minimal or zero - and that's beginning to shake the golden throne that folk like Mark Zuckerberg and Sam Altman have placed themselves on. If investors suddenly see through that emperor's invisible cloak, and see the reality of what they're paying for - a subscription service to an 'employee' that can't perform its duty - then they will bail on it, and start rehiring low-paid human beings to do the same task - humans who are properly capable of conscious thought, problem solving skills and communication that clever AI bots can't manage. 

Those suffering from the worst kinds of "AI Psychosis" (which is a very real condition that mental health practitioners are now having to cope with) keep coming back with the defence that AI will save us, that AI really will be the saviour of humanity. It will cure cancer, it will fix climate change, and it will allow those of us with mundane jobs to shrug all that off and take up a life of creative leisure. Some folk truly believe that AI has already attained consciousness, but it's doing nothing more than performing a few cute tricks to fool humans into thinking it's behaving like a human. AI bots are showing signs of racism purely because they are fed with an endless diet of crap from social media where racism is rife, so naturally it will try to 'please its masters' by behaving like them in the same way a kid will copy their mum or dad when they're smoking a cigarette, holding up a pencil as a fake ciggie (before taking up the real thing probably in the not too distant future). 

I watch our institution as it spirals into an endless morass of bullshit. We are currently moving through the process of purchasing yet another AI framework, we are contemplating paying Microsoft through the nose for more Copilot stuff even though our storage is almost gone, we still have massive problems with privacy, governance, data breaches and abusive user behaviour and yet we throw money at the wrong things time and time again. 

I feel more and more like I need to get out of this gig before it's too late.