Pity the policymakers in the AI jobs tsunami

{"text":[[{"start":null,"text":"

"}],[{"start":7.29,"text":"You’re standing on the beach in the dark, trying to guess whether the next wave, which you can already hear, is going to sweep everyone off their feet, or peter out and merely tickle your toes. This must be roughly how it feels to be a policymaker this year — deeply unsure about what is coming for the labour market, let alone what to do about it."}],[{"start":34.980000000000004,"text":"Kristalina Georgieva, managing director of the IMF, has said that artificial intelligence is “already like a tsunami hitting the labour market”. Ford chief executive Jim Farley thinks AI is going to replace “literally half of all white-collar workers” in the US."}],[{"start":57.47,"text":"And yet there are counter-signals too: a study of 25,000 workers and 7,000 workplaces in Denmark has found that rapid corporate adoption of generative AI chatbots in 2023 and 2024 had “no significant impact on earnings or recorded hours in any occupation” and “minimal” economic impacts overall. Even within professions, the signs are mixed. Generative AI seems well suited to replacing the work of junior lawyers, for example. Yet salaries for newly qualified lawyers are surging. You don’t have to be an economist to know that the price of something doesn’t usually go up when demand for it goes down."}],[{"start":111.86,"text":"What is a policymaker to do? One temptation might be to wait and see what happens, then try to help the people who are knocked off their feet. When automation and globalisation hollowed out factory and mining towns between the 1970s and 2000s, the most popular policy response was to offer “retraining” to those who lost their jobs. But many of these programmes failed to deliver, especially when they tried to push people into the supposed “jobs of the future”. In particular, the “coal to code” boot camps, which promised to retrain unemployed coal miners in the US as entry-level computer programmers, have not aged well. "}],[{"start":162.9,"text":"To be fair, it was not unreasonable to presume a decade ago that demand for coding skills would remain robust, no matter what else happened. After all, the first industrial revolution replaced many skilled craft workers with machines, but created new demand for people who could build and maintain those machines. According to the historian EP Thompson, as late as 1818, The Book of English Trades did not even list the roles of engineer, steam-engine maker, or boilermaker. Yet just 10 years later, a directory called The Operative Mechanic and British Machinist was published that ran to 900 pages."}],[{"start":214.73000000000002,"text":"But this time around, the machines appear to be displacing their own makers — or, at least, their makers’ apprentices. Analysis by Oxford Economics suggests that much of the recent rise in graduate unemployment in the US is driven by the struggle of newly minted computer science graduates to find jobs. While employment for over-27s in computer science and mathematical occupations has risen 0.8 per cent since 2022, employment for those aged 22-27, or recent graduates, has declined by 8 per cent. Coding boot camps are also beginning to close."}],[{"start":263.03000000000003,"text":"Rather than policymakers having a guess, workers themselves are probably best placed to see the ripples of danger and opportunity in their own sectors and professions. A better policy response, then, would be to help people take matters into their own hands."}],[{"start":283.98,"text":"One example comes from Sweden, which introduced a sort of furlough scheme for life-long learning in 2022. Under an agreement between employers, unions and the government, workers can take time off from their jobs to train in something new, while being paid 80 per cent of their salary (up to a limit)."}],[{"start":309.21000000000004,"text":"The policy is very much in keeping with the Swedish approach to labour market disruption, which is to “protect people, not jobs”. In other words: creative destruction is necessary, but people shouldn’t pay the price."}],[{"start":330.66,"text":"The Swedish policy has proved popular with people who want to move up to better work, or out of harm’s way. There have been 136,000 applications so far, according to Thomas Carlén from the Swedish Trade Union Confederation. He said there had been a mixture of blue-collar and white-collar applicants, 80 per cent of whom were employed. One condition of the scheme is that applicants have to show their chosen training will “strengthen their position in the labour market”. While this still requires some guesswork, it is better than restricting people to certain pre-chosen courses."}],[{"start":372.95000000000005,"text":"What else could policymakers do proactively? It would be wise to shore up the social safety net for a world of work in which more people go it alone, either through choice or necessity. Young people unable to find a foothold in large companies might find it easier and cheaper these days to start a business of their own, thanks to AI coding and marketing tools. Experienced workers pushed out by employers keen on automation might choose to set up boutique competitors that promise the human touch. But the self-employed often slip through holes in social security systems, especially when it comes to pensions."}],[{"start":422.53000000000003,"text":"When you don’t know what’s coming next, the best policy is neither to become paralysed, nor to become prescriptive. Better to give people the support and the confidence to find their own way through. Even if the AI “tsunami” turns out to be an ankle-depth anticlimax, policymakers won’t regret making sure that everyone can swim."}],[{"start":460.59000000000003,"text":""}]],"url":"https://audio.ftmailbox.cn/album/a_1752656508_6318.mp3"}