Artificial intelligence replaces entry-level employees whose tasks can be performed by generative AI tools such as Chatgpt, finds a rigorous new study.
Employees in the early career in areas that are most exposed to AI have experienced a decrease in employment of 13% since 2022, compared to more experienced employees in the same areas and measured against people in sectors who are less buffet due to the rapid technology, according to a recent labor paper By Stanford economists Erik Bynjolfsson, Bharat Chandar and Ruyu Chen.
The study contributes to the growing amount of research It suggests that the spread of generative AI in the workplace will probably disrupt the labor market, especially for younger employees, the authors of the report said.
“These great language models are trained in books, articles and written material on the internet and elsewhere,” Bynjolfsson told CBS Moneywatch. “That is the kind of book to learn that many people become at universities before they enter the labor market, so there is a lot of overlap between these LLMS and the knowledge that young people have.”
The research in particular emphasizes two fields where AI already seems to replace a considerable number of young employees: software engineering and customer service. Between the end of 2022 and July 2025, employment at entry level in those areas fell by around 20%, according to the report, while employment for older employees grew in the same jobs.
In general, employment for employees aged 22 to 25 in the most AI-bloted sectors fell by 6% during the study period. For comparison:, according to the researchers, employment in those areas increased between 6% and 9% for older employees.
The analysis reveals a similar pattern that takes place in the following fields:
- Accounting and auditing
- Secretarial and administrative work
- Computer programming
- Sale
Older employees, who have generally navigated for a longer period in the workplace, have rather the chance of the types of communication and other “soft” skills that are more difficult to teach and that employers are reluctant to replace AI, suggest the data.
“Older employees have a lot of tacit knowledge because they learn tricks from trade from experience that may never be written down somewhere,” Brynjolfsson explained. “They have knowledge that is not in the LLMS, so they are not replaced so much by them.”
The study is unusually robust in view of the fact that generative AI technologies are only a few years old, while experts are just starting to dig in the impact on the labor market. The Stanford researchers used data from ADP, which offers wage processing services to employers with a combined 25 million employees, to keep track of employment changes for full-time employees in professions that are either more or less exposed to AI. The data contains detailed information about employees, including their age, and precise job titles.
AI is not only in danger of taking jobs away from employees. Just as with earlier cycles of innovation, it will die out some jobs while creating others, said Bynjolfsson.
“Tech has always destroyed jobs and creates jobs. There has always been this turnover,” he said. “There is a transition over time, and that is what we see now.”
Enlarged or automated?
In areas such as Nurse AI, for example, people are more inclined to increase human employees by taking Rote tasks, according to the proponents of technology, spending more time on patients on patients.
Although the work at the entry level has fallen in professions that are most exposed to AI, such a decline has not taken place in jobs where employers want to use these tools to support and expand what employees do.
“Employees who use these tools to increase their work benefit,” said Bynjolfsson. “So there is a recasting of the type of employment in the economy.”
Advice for young employees
According to Brynolfsson, employees who can learn to use AI to help them do their job better will be best positioned on success on today’s labor market.
A recent report AI personnel file Burtch Works showed that starting salaries for AI employees at entry level increased by 12% from 2024 to 2025.
“Young employees who learn how to use AI can be used effectively can be much more productive. But if you just do things that AI can do for you, you don’t have much added value,” Bynjolfsson told CBS Moneywatch.
“This is the first time that we get clearer proof of this type of employment effects, but it is probably not the last time,” he added. “It is something we have to pay more and more attention to when it develops and companies learn to benefit from things that exist.”
#study #sheds #light #kind #employees #losing #jobs