New study: Some early-career workers are losing their jobs to AI

Some early-career workers are losing their jobs to AI: study

Is artificial intelligence coming for your job? A new study takes a look at how AI is already impacting different types of workers. FOX 5's Josh Rosenthal reports.

WASHINGTON - Is artificial intelligence coming for your job? A new study takes a look at how AI is already impacting different types of workers.

What we know:

A new working paper by Stanford economists takes a look at changes in the labor market for jobs exposed to generative artificial intelligence (AI).

Among the findings is that early-career workers (ages 22-25) in the most AI-exposed occupations, like coders and people who work in call centers, have experienced a roughly 13 percent decline in employment since the widespread adoption of generative AI, including large language models (LLMs) like ChatGPT.

"That kind of double-digit decline was really striking and actually a little unexpected to me, how stark it was," said Stanford Digital Economy Lab Director Erik Brynjolfsson, one of the study's authors.

He added that employment for more experienced workers in the same occupations has remained stable or continued to grow.

"The folks who are on the job, they get a lot of tacit knowledge, tricks of the trade that they learn that’s never been written down before. Well, the LLMs, they mostly absorb a lot of written language, and so there may be a little bit more of an overlap for stuff that people learn from reading, whereas the tacit knowledge is something that the LLMs are less likely to have as much of," Brynjolfsson explained.

Big picture view:

The news isn't all doom and gloom. Brynjolfsson added that another of the paper's findings is that the use of AI to augment work actually led to faster employment growth than average.

"Some people go to LLMs, go to these large language models, and ask them, ‘hey, teach me how to do this new skill. Show me how to make a better presentation. Help me learn something.’ When people were using it that way, we actually saw an increase in employment."

Dig deeper:

The full working paper can be found here.