Your job's safe for now! AI just isn't quite ready to be unleashed on the sceptical world

June 22, 2024
Temple Melville

IN the old days you simply whispered 'Cortina' and everyone knew that you knew. Now, mention AI and the same thing happens. Everyone wants to know what are you doing about it and have you lost your job yet?

In the same way that a few years ago that a decidedly grubby apple juice producer mentioned it was going to use blockchain and promptly had its share price disappear in the clouds, the same thing is happening today with AI. You are rewarded if you mention it and punished if you don’t.

Apple announced some minor AI enhancements and its share price rocketed. It makes no sense as the increased earnings (if any) will be miniscule relative to the rest of Apple’s business. Investment analysts around the world are beginning to realise that AI and advancing it is a very expensive process, not only non-profit making but, as one put it, a very expensive PR exercise.

For those of a sceptical nature, the Shiller metric has reared its head again. It basically describes how “frothy” the markets are. If you have been paying attention you will know we now have meme coins and businesses (all with literally no value) reaching for the stars. When was it last this frothy? 1999 just before the dotcom bust. And before that? 1929 – and if I have to tell you about that you probably shouldn’t be here.

The present froth is all about how AI will take over lots of jobs so increasing productivity and profits. But all the evidence so far points almost in the opposite direction.

Apparently only 2% of Brits are using ChatGPT, as they see a mismatch in what it produces and the eventual results. Not least, reading anything generated by AI is deadly boring with no spark. How could it be otherwise? What it produces – or should I say reproduces – ends up as extremely anodyne pap. Never mind the glaring errors.

Apparently Google recently postulated that you could stay healthy by eating one small rock a day. The best AI seems to be generating lots of stuff we don’t need, and is having a hard time automating anything useful.

The AI phenomenon appears to be promulgating hallucinations and group think – so excoriated by psychologists. After all, if the man next door tells you he is installing AI and it is great (and yes we still have to train it properly) then unless you are very strong willed, and because you keep reading similar statements, it is very hard to resist.

There are lots of examples like the small rock one above, but I particularly liked the ChatGPT rogue period which spoke in Spanglish. It was almost certainly more interesting than what it usually produced.

In today’s self-reinforcing tech world, it is a brave man who stands against the tide. Any manager, besieged by his IT department demanding AI will almost certainly cave in. It is rather akin to believing in God. You might not, but it’s probably a good idea to hedge your bets. As Shiller says in his book - 'Irrational Exuberance' - these sort of events lead money and investment away from important things like infrastructure and education and into dead ends with no beneficial outcome.

What does rather please me about it is that all the students who thought they didn’t need to study have had a rude awakening. Quite apart from the similarity of answers, the answers themselves tend to be either rubbish or self-evident – not what University Professors look for.

Swathes of dissertations have been junked and students failed, particularly in the States. And in an update from McDonalds – who have “thoughtfully” removed AI from their ordering systems after it proved extremely costly to fix the wrong deliveries (it appears their system really likes ice cream topped with bacon). What that says about the people actually handing out the orders is a subject for a psychology dissertation all on its own.

You can probably tell this article was created without using AI. Apart from anything else the English is correct.