Following the amazing recent advances in artificial intelligence, many people fear ending up in the economic dustbin. On Google, global searches like “Is my job in danger?” They have doubled in recent months due to many fearing that they will be replaced by large language models (GML). There are signs that big, disruptive changes are on the way. In a recent article, OpenAI’s Tyna Eloundou and colleagues state that “around 80% of the US workforce could see at least 10% of their work tasks affected by the introduction of GML”. Another study indicates that legal services, accounting and travel agencies will face an unprecedented upheaval.
Yet economists are more into making predictions about automation than testing them. In the early 2010s, many predicted with great fanfare that robots would kill millions of jobs, but then fell silent as employment rates in the rich world reached all-time highs. Among those catastrophists, few have a good explanation for why countries with the highest rates of technology use (such as Japan, Singapore, and South Korea) are consistently among those with the lowest unemployment rates.
We present here our first attempt at the jobs. Using US data on employment by occupation, we have looked at white-collar workers. They are people who work in all kinds of sectors, from administrative support and financial operations to copywriting. White-collar jobs are believed to be especially vulnerable to generative artificial intelligence, which is getting better and better at logical reasoning and creativity.
However, there is still little evidence of an effect of artificial intelligence on employment. In the spring of 2020, the share of white-collar workers in total employees increased, as many people working in the service sector lost their jobs when the covid-19 pandemic broke out (see graph). The proportion is lower today, because leisure and hospitality have recovered. However, in the last year the proportion of employment in professions supposedly at risk from generative artificial intelligence has increased by half a percentage point.
Of course, it is still early. Few companies are still using generative artificial intelligence tools on a large scale, so the consequences for employment could be yet to come. However, another possibility is that these new technologies end up destroying only a small number of jobs. Artificial intelligence can be efficient at some tasks, but may not be so efficient at others, such as managing and identifying what others need.
Artificial intelligence could even have a positive effect on employment. If the workers who use it become more efficient, your company’s profits will increase, allowing bosses to increase hiring. A recent survey by Experis, a recruitment company specializing in the field of information technology, points to this possibility. More than half of British employers expect AI-related technologies to have a positive impact on their workforce in the next two years.
To see how things play out, we’ll be posting updates to this review every few months. For now, though, the workplace apocalypse seems far away.
© 2023 The Economist Newspaper Limited. All rights reserved
Translation: Juan Gabriel López Guix
