The AI jobpocalypse has begun.
For professions at risk of being replaced by AI, such as software engineering and customer service, entry-level employment has fallen 13% since 2022, a Stanford study has found.
The research, which analyzed millions of jobs using data from payroll processing firm ADP, painted an especially grim picture for tech. Employment for software developers ages 22 to 25 has fallen by almost 20% since ChatGPT’s public release.
To hear it from Anthropic CEO Dario Amodei, this is merely the start. AI could wipe out half of entry-level white-collar jobs over the next five years, he told Axios.
Software engineers are particularly at risk because of tools developed by Amodei himself. Stanford’s AI Index Report said that in 2023, the top-performing AI system could solve 4.4% of software engineering problems measured; by 2024, OpenAI’s o3 model reached more than 70%.
It’s a dismal situation for computer science graduates, who were once promised six-figure salaries and cushy lifestyles right out of school. Now, even a college degree isn’t a guaranteed leg up in the job hunt: The study found that professions with a high percentage of college grads had declining employment, while those with a low percentage saw rising employment.
“It used to be the land of milk and honey, where students had considerable choice,” Paul Glanting of San Francisco State University’s business school told The Standard in May. “But that abundance just isn’t there anymore.”
Opportunities at cushy tech firms have dwindled. Pre-pandemic, new grads made up 15% of hires in Big Tech; now it’s 7%, according to a report from VC firm SignalFire.
Startups aren’t picking up the slack. Seed-stage companies had 21% fewer employees in the first half of 2025 than they did in the first half of 2020, according to a report from Carta. These founders are “leveraging AI to accomplish much more with small staffs than was previously possible,” the report said.
It’s not all doom and gloom, however. New grads who studied machine learning can make seven figures at companies like data analytics firm Databricks, according to a Wall Street Journal report. And the Stanford study found that for experienced software engineers or customer service experts, employment opportunities remained stable or even grew. Some non-tech industries seem AI-proof (for now), like health aides and front-line production and operations supervisors.
With these shifts, Emily Schaffer, an executive at YUPRO Placement, which trains historically overlooked talent for the job market, has been forced to change her approach.
Instead of training people for office jobs, she is looking to industries that will be less affected by AI, such as property management.
But she cautioned that companies need to think carefully before slashing entry-level jobs. “You cannot get a mid-career workforce unless you also have an early-career workforce,” she said.