Saturday, June 24, 2023

Point of Order: The good thing about artificial intelligence is what it might do for human intelligence

As the man who seized Boris’s crown, Britain’s PM Rishi Sunak has his work cut out.

But his latest brainwave – making the UK a global centre for regulation of the new wave of artificial intelligence (AI) – is unlikely to help.

While recent times suggest that more detailed prescription is not the way to fruitful innovation, the confidence of politicians, bureaucrats and general experts has not diminished.

Witness the astonishing outburst at the so-far modest appearance of ChatGPT and other AI chatbots aka large language models.

Given the world’s many problems, is it not a trifle hysterical to obsess over whether AI will kill us all (i.e., the Skynet scenario)?

Fortunately, Rishi does not think that we’re doomed.

But it’s more than a little depressing that he and his advisers hardly seem to have stopped to reflect on their (indeed our) lack of knowledge of the future; and ponder the historical unwisdom of trying to steer a process with extraordinary possibilities that has barely started.

Remember the European Union tried to get ahead of the tech play through the application of pure reason with its Orwellian-sounding General Data Protection Regulation (GDPR) – and succeeded in pushing the continent even further behind in tech development.

So it might be useful to listen to someone who does reflect deeply on these things – someone like futurist Tyler Cowen (well, an economics professor in day-to-day life).

His snapshot model for understanding AI potential is that $20 per month currently gets you unrestricted access to a high quality research assistant, architect and colleague.

If you can teach yourself to make full use of these, you gain a quite remarkable ability to productively leverage your own knowledge.

And if you can’t or won’t, your productivity will decline relative to those who can.

This would seem to be transformative.

Dependence on expensive knowledge experts is likely to be greatly reduced (think of the possible impact in medicine, law and architecture, to name a few areas).

Which in turn ought to make other forms of expertise – like physical manipulation of things, and execution of activities – relatively more valuable.

While tangible assets complementary to new patterns of activity (such as electricity and well-located property, Cowen speculates) will be more sought after.

You can foresee less need for government; unless, perhaps, you hold to the adage that rules are needed for stupid people.

While this may offer the best possibility for improving our quality of life, it might also make politicians – well today’s politicians – uneasy.

Because the enormous long-run benefits would come with short-run creative disruption. And not a few people would find their skills and status devalued.

This would be a fundamental challenge to current cramped and conservative policy settings, which posit that we all keep our place in the slowly shuffling-forward queue, with Jacinda exhorting the back rows to bunch up a bit more.

And that no one should endure bad luck or change outside their control.

The spiritual home of this policy is surely the European Union. And there seems to be a price. Its economy has fallen decisively behind that of the United States, frets Gideon Rachman in the Financial Times. It was 10% larger in 2008, now 25% smaller on the equivalent measure.

And yet this is the very place which needs the creative disruption and higher productivity of AI to pay for the pensions of an ageing population; for levelling-up to ‘end’ child poverty; and to achieve net zero decarbonisation.

Rishi Sunak is right to believe that AI is likely to be consequential and that gives his country great opportunities.

But he’s much less sound if he really believes that current regulatory models – and particularly the international ones – are the key to exploiting those successfully.

Point of Order is a blog focused on politics and the economy run by veteran newspaper reporters Bob Edlin and Ian Templeton

No comments: