Godwin’s law asserts that any online discussion eventually ends up invoking Hitler. There’ll soon need to be something similar for AI. Can’t write that business plan? Use AI. Can’t get a girlfriend? Use AI. It encourages us to give up on all that is challenging, but worthwhile.
A few years ago, I was with a senior politician at a big media conference. Looking through the agenda, she rolled her eyes. “AI, AI, AI – it’s so boring.” Maybe that’s an indictment of the failure of the political class to think seriously about big issues. But I can’t help but be terribly sympathetic to that sentiment.
I write this as a ChatGPT subscriber and extensive user. I just asked it to find out the name of that Hitler law in the first line. Nor do I resent the field’s leading lights. I think they are predominantly geniuses with good intentions.
Anti-human instinct
I’m sceptical of their more grandiose claims but only in the same way that I’m a climate change sceptic. I couldn’t debate a scientist on the topic. I haven’t analysed temperatures and can’t make much sense of graphs. I believe what I’d like to be right. The term Climate Emergency repels me because I don’t like the implications. I don’t want humans to stop making stuff or going on holidays. I think humans are great, not a problem to be solved. So I hope that the science is overhyped or that techno-optimists are right. That our ingenuity solves the problem, as it always has.
I feel the same instinctive hostility toward the anti-human strain of AI. It promotes such a reductionist view of our capabilities. Predictions of a job apocalypse encourage that, seeing humans as easily replaceable. (Ian Leslie‘s 5 Reasons There Won’t Be an AI Jobs Apocalypse is an excellent antidote to this thinking).
I don’t like the shysters who prey upon this fear
Everyone will be out of a job by 2029. You need to start accumulating capital now to survive. Join my masterclass to find out how!
Or proselytise that salvation lies only in prompting. Become the Dr Doolittle of the machine world and stay relevant. Don’t bother with bettering yourself. Just learn how to instruct machines.
Self-fulfilling prophecy
This AI doomerism is a self-fulfilling prophecy because:
- We’re using AI tools most prolifically in areas where humans should thrive, namely communication. To pretend you’re interested in someone’s email or for drafting content (the more sophisticated at least delete the giveaway em dashes). But companies will eventually realise days spent exchanging AI generated emails aren’t delivering great productivity gains. And that, professionally worded, but generic content doesn’t stand out against thousands of identikits. If I drafted corporate press releases, I would feel threatened that AI is perfectly capable of generating the same corporate waffle. So perhaps this is the moment to use fewer empty phrases and be a bit more succinct and original. It’s a big part of the reason populist politicians cut through. Trump’s comments on Israel and Iran – they don’t know what the fuck they’re doing – landed a lot better than Starmer’s banal talk of restraint, calm and a return to diplomacy.
- AI is not only making you more expendable but stupider too. A recent MIT study shows it is attacking us by stealth. Increasing reliance rots the brain. Tracking participants for four months, LLM users consistently underperformed at neural, linguistic and behavioural levels.
Reading
MIT allowed that group to outsource their research and writing to LLM tools. Commenting on the study in The Times, James Marriott wrote:
Reading and writing are the cornerstones of thought: serious reading, I suspect, is the one habit that unites virtually every man and woman of genius who has ever lived. It may turn out that to abolish reading and writing is almost to abolish human genius.
He argues that reading and writing form the basis of logical argument. They provide the skills that enable us to make informed judgements. It’s why, for instance, many of the world’s top fund managers come from humanities backgrounds. Or why Warren Buffett spends six hours a day reading. It teaches them to synthesise a large amount of information and form a larger thesis on how that plays out in the markets.
The writer Ezra Klein explains the value of reading wonderfully below. It’s not importing information into your head. Rather, reading is a two-way activity. It changes you too. It forms connections you didn’t have before. And it also establishes what Ian Leslie calls cognitive endurance – a learned ability to really engage with a topic
An excellent servant but terrible master
As I was once warned about alcohol, AI is an excellent servant, but a terrible master. My new year’s resolution was to read more. And I’ve found ChatGPT a wonderful aid in that – to look up unfamiliar concepts or provide wider context. But it can’t replace the reading itself.
AI is boring when it sees humans as a problem. It makes life more mundance. It diminishes authenticiy and originality and prioritises outsourcing over doing things better. Like climate catastrophists, its message is too often one of human worthlessness. And we risk fulfilling its gloomier prophecies if we absorb that narrative.
