It’s too early to say about AI but….. – Adam Smith Institute

By Tim Worstall

If the roll out of AI does kill off a lot of jobs then this is the way it will, even should, happen.

‘They’ve cut all entry-level jobs’: unlikely seaside town on frontline of AI threat

We should note that AI will only work if it does kill off a lot of jobs. That is our measure of success: we can now gain this production without having to use human labour to get it. Thus there is more human labour that can be devoted to solving some other human problem, want or desire. We are richer by the value we place upon that new solution. After all, jobs are a cost, not a benefit.

But the how:

As bootmaking mechanized, 153,000 jobs disappeared, and 140,000 emerged. Incumbents mostly stayed – the trade collapsed primarily because young people stopped entering.

As we like to point out the tail end of a technology lasts a long, long, time. The London Hydraulic Power Company was still energising factories a good 50 years after electricity was the new technology of choice. Might as well sweat extant assets for as long as it is possible sweat them, no? So with human capital. The technological change can simply be that there are no entrants into training and the workforce shrinks by that natural process of ageing.

Of course, the change can be faster than that leading to active displacement. But it’s still true that the first – and most reasonable – step is to stop taking new labour in and trying to train it up in the dying occupation.

If AI does lead to the reduction of certain forms of employment – we’re willing to be open on that as yet – then this is the way that it should be done. Don’t hire any new people.