AI crushes the labor market at the start of a career, according to Stanford’s study

If you suspected that the AI ​​removes jobs from young workers, there is now data to support this.

Three economists in the Digital Economy Laboratory of the University of Stanford – Professor Erik Brynjolfsson, researcher Ruyu Chen and the postdoctoral assessment Bharat Chandar – published a paper On Tuesday, this found that workers at the start of their careers aged 22 to 25 in the jobs most exposed to AI “experienced a relative 13% drop in employment”.

“On the other hand, employment for workers in less exposed fields and more experienced workers in the same professions has remained stable or continued to grow,” wrote the researchers.

In fact, for professions that cannot be easily replaced by AI, such as home aid, employment possibilities for young workers seemed to increase faster than for elderly workers.

The effect was visible even when we take into account the specific shocks of the company and other potential causes such as changes in remote labor policies, the effects of the pandemic on the education system, the slowdown in technological hiring or the trends in cyclic employment, noted the researchers.

“The AI ​​revolution is starting to have a significant and disproportionate impact on entry -level workers on the American labor market,” said the researchers.

The results are saved by anecdotal evidence that has accumulated for months.

CEOs in all industries have been opened on their expectations–and their Business policies already in action–For artificial intelligence to manage the work that some new employees would have otherwise.

“There is a real fear that I have an entire cohort, those which are graduated during the early transition of AI, could be a lost generation, unless the standards of education, education and hiring, are adjusted,” said John McCarthy, aggregate professor of global work and work at the school of industrial relations and work of Cornell, said to Gizmodo Earlier this month.

But while some experts had sounded alarms, others had hesitated to point their finger on the AI ​​without tangible data.

This is why the Stanford newspaper is significant. This is a kind study of the first and this shows that data that can support a trend of young graduates complained and worried for months: this AI comes well for their work.

Older workers are spared

The researchers compared the changes in the data on the employment of the end of 2022 in mid-2010, graceful of the ADP payroll processing company, which is one of the largest in the United States and represents more than 25 million workers.

The results have shown that industries that have largely adopted AI, such as software engineering, have shown a significant decrease in jobs available for young graduates after 2022.

While employment has dropped for young graduates looking for work in AI -supported industries, researchers found that older and more experienced workers were largely spared.

While workers aged 22 to 25 have had a drop in employment since 2022, employment for elderly workers aged 35 to 49 increased, according to researchers.

This may be due to the fact that AI is good in basic tasks, that than a recent graduate with less practical work experience than an older worker should manage.

But even if the automation of these basic tasks seems to be a good commercial strategy, this kind of work at the start of their career is crucial for the training of the next generation of the workforce. If these training possibilities are not offered to entry-level workers, the future of the workforce does not seem to be unrecognizable.

“I fear that the current generational pressure can evolve towards a permanent reconfiguration of career paths at the start of his career,” McCarthy told Gizmodo earlier this month. “There is a real fear that I have an entire cohort, those which are graduated during the early transition of AI, can be a lost generation, unless politics, education and hiring adapt.”

Automation vs increase

Within the industries with a strong adoption of AI, that companies intend to use AI to automate or increase human work have made a huge difference, according to the newspaper.

The employment drops were widely concentrated in jobs where AI was used to completely or partially replace the workloads of certain employees, rather than to complete it.

In a Previous June documentThe co-author Brynjolfsson argued that AI companies should develop benchmarks that test the way AI models can collaborate with humans to jointly resolve tasks, rather than relying solely on existing benchmarks that assess AI in the absence of humans. This can help move the objective of integrating AI from automation to increase and collaboration, Brynjolfsson and its co-author on the June Andreas Haupt document.

AI is under development as an automation tool above all at the moment, but the results suggest that this may not be its best use if we want AI to be a tool for a positive change.

The AI ​​could help individual workers by attenuating the burden of heavy workloads while continuing to generate productivity gains. Or it can be used to fully automate certain jobs, moving away from the possibilities of young careers from young graduates who are supposed to constitute the foundations of a well-trained future workforce. Which of these results will be reality will be finally determined by the way in which the business world decides to develop this revolutionary technology in the future.