Adapting to AI: how workers can manage disruption

The launch of ChatGPT has reinvigorated the debate about the impact of technology on the labor market.

We’ve become accustomed to hearing doomsday predictions about the rise of generative AI and the various job losses likely to result from it.

For instance, Goldman Sachs warned that a massive 300 million jobs could be at risk. Meanwhile, influential figures like Steve Wozniak and Elon Musk urged caution in the rapid development of artificial intelligence (except for self-driving cars, obviously).

Indeed, Sam Altman, the leader of OpenAI, admitted to feeling a bit worried about AI's future. This concern is shared by Ilya Sutskever, OpenAI's top scientist, who said that it might become easy to cause significant harm with AI. OpenAI's own report also hinted that most jobs could be in some danger due to their technology.

Unwarranted concern?

Concerns around the future of jobs hark back to the worry when Oxford academics Frey and Osborne published their 2013 paper on how AI might affect employment. Surprisingly, those anxious predictions haven't materialized, and we're actually experiencing historically low unemployment rates.

A recent report from CompTIA reveals that the technology industry saw a significant increase in jobs last year across all states. The sector saw a net growth of 3.2%, leading to the creation of over 280,000 jobs nationwide. These encouraging numbers show that the technology industry continues to expand, making it a crucial contributor to the U.S. job market.

The report also shows that the technology sector employs a whopping 9.1 million people in the United States, covering both technical and non-technical roles. These results underscore the enduring strength of the technology industry across different parts of the United States.

Making an impact

That’s not to say that technological disruption doesn’t have an impact, of course, but it’s important to keep things in a degree of perspective. Indeed, research from the London School of Economics explored how technology impacts the livelihoods of workers, and the consequences were relatively minor.

The researchers examined the impact of technology on the jobs and income of Swedish workers. Their findings show that the possible drawbacks are quite small, resulting in a 2-5% drop in earnings and a 1-2% decrease in job opportunities. However, it's important to note that these losses hit harder for those who earn less.

This fairly modest impact was primarily because those workers affected by technology were able to migrate into different professions. It also suggests a reasonable degree of flexibility in our career choices, thus allowing people to compensate for the drop in their initial income by transitioning into fields less affected by technology.

The importance of flexibility might also explain why those in the lowest income brackets suffered the most from new technologies. These people often experienced a significant fall in their career earnings, with the average worker seeing their income fall by 8-11%. While they were less likely to stick to their original jobs compared to higher earners, they also faced challenges in finding new opportunities, especially ones that paid well.

The key role that job flexibility plays in our ability to weather the storm of technology was underpinned by the finding that a substantial portion of lost income was linked to periods of unemployment, accounting for about a third of all employment losses, as opposed to retraining, which contributed to less than 10%.

Adapting to change

So how can people make such a transition, and how can society help them? While the risk of technological disruption to the labor market tends to grab most of the headlines, offshoring is at least as big a risk, if not more so, and this perhaps provides a degree of guidance on how society should help people to adapt.

Research from Chicago Booth explores how countries can help people whose livelihoods are affected by offshoring. The researchers analyzed administrative microdata at the individual worker level. It allowed them to link each worker with any training programs they had been assigned and their subsequent labor-market outcomes.

The researchers monitored each worker for over ten years, with two key types of training assessed. The first was a classroom-based program designed to help build human capital, with participants earning certificates in various new skills.

The second intervention was job training, with apprenticeship-type approaches subsidized by the government. The aim was to get people’s foot in the door and allow them to build from there.

The analysis found that classroom training did help people find new work, with the typical worker getting an extra 25 hours each month a few years after participating in the program. This is largely because knowledge-based reskilling helped people to move into new occupations.

The researchers explain that this underlines the value of training in the labor market, with participants greatly helped in their job search. This was especially true for those who faced structural challenges in the labor market, such as their jobs being offshored.

The on-the-job training programs weren’t as effective, however, with the benefits tending to fade once the wage subsidies eased off. As a result, the researchers believe that classroom training is probably the best approach for supporting workers whose jobs are affected by offshoring or automation.

The study demonstrates that investing in the skills and knowledge of displaced workers is a smart move. It can transform someone who would have relied on government benefits into a taxpayer who contributes to society. While it remains to be seen if generative AI will be as disruptive as its advocates claim, we do at least appear to have an idea of how to adapt to it.