Artificial Intelligence is replacing the entry-level workers, whose jobs can be done by generative AI tools such as flat, get a rigorous new study.
Initial career employees have experienced a 13% decline in employment since 2022 in the most exposed areas for AI, compared to more experienced workers in the same areas and when measured against those who have recently emerged by a rapid emerging technology, recently. working Paper Stanford economists Eric Brianjolphson, Bharat Chandra and Ruu Chen.
Adds to the study Rising body of research The authors of the report said that the dissemination of generic AI in the workplace is likely to disrupt the job market, especially for young workers, the authors of the report said.
“These big language models are trained on books, articles and written materials found in the Internet and other places,” Bryngolphson told CBS Manivatch. “This is the type of book that many people get in universities before learning, before they enter the job market, so there is a lot of overlap between these LLM and young people of knowledge.”
Research especially highlights two areas where AI already appears to be pressed by pressing a significant number of young workers: software engineering and customer service. At the end of 2022 and July 2025, according to the report, entry-level employment in those areas declined by about 20%, while in equal jobs, employment for old workers increased.
Overall, the most exposed areas fell 6% during the employment study period for workers between 22 and 25 years of age. Comparatively, according to researchers, employment for older workers increased between 6% and 9%.
Analysis reveals a similar pattern in the following areas:
- Accounting and accounting
- Secretary and administrative function
- Computer programming
- Sales
Older employees, who have generally navigated the workplace for a long time, have more likely to raise the type of communication and other “soft” skills that are difficult to teach and can be reluctant to replace with employers AI, suggest data.
“Older workers have a lot of silent knowledge because they learn the tricks of business from experience that can never be written anywhere,” Bryncolphson explained. “They have knowledge that is not in LLM, so they are not being replaced by them.”
The study is abnormally strong, given that generative AI technologies are only a few years old, while experts are only starting to systematically digging in effects on the labor market. Researchers at Stanford used data from ADP, which provides parole processing services to employers with a combined 25 million workers, which is to track employment changes for full -time workers for AI or more or less businesses. The data included detailed information about workers, including their age and accurate job titles.
AI does not only threaten to quit jobs from workers. With the previous cycles of innovation, it would extinct some jobs while creating others, Bryngolphson said.
“Tech has always been destroying jobs and building jobs. It has always been a turnover,” he said. “There is an infection over time, and that is now we are watching.”
Promoted or automated?
For example, according to technology supporters in areas such as nursing AI, health care doctors are likely to increase human workers to spend more time to focus on patients.
While entry-level employment has fallen into businesses that are in touch with AI, there is no such decline in jobs where employers want to use these devices and use these devices to use these devices that employees do.
“The activists who are using these devices to increase their work are benefiting,” Bryngolphson said. “So there is a rearrangement of the type of employment in the economy.”
Advice for young workers
Workers who can learn to use AI so that they can help them do better work, will be best deployed for success in today’s labor market.
recently Report The AI ​​staffing firm Burtch Works found that introduction-level AI workers increased by 12% from 2024 to 2025.
“Young workers who learn how to use AI effectively can be very productive. But if you are just doing things that AI can already do for you, you will not have that much value,” Bryngolphson told CBS Manivatch.
He said, “This is the first time we are getting clear evidence of such employment effects, but it is probably not the last time,” he said. “This is something that we need to pay attention to because it develops and companies learn to take advantage of things that are out of there.”
