AI and junior developers: what the data shows, and how to stay employable

Software development is one of the jobs most exposed to AI, and early-career developers feel it first.

What the data shows

Stanford Digital Economy Lab researchers Erik Brynjolfsson, Bharat Chandar and Ruyu Chen studied payroll data across AI-exposed occupations, including software development. In the August 2026 revision of their paper, they report that employment of workers aged 22 to 25 in those occupations is 19% below where it would be had it kept pace with less-exposed peers. More experienced workers in the same jobs were not hit the same way.

At the same time, developers have adopted AI tools fast. In Stack Overflow's 2025 Developer Survey, 84% of respondents said they use or plan to use AI tools in development, up from 76% the year before, and about half of professional developers use them daily.

But trust is low. 46% of respondents said they distrust the accuracy of AI output, against 33% who trust it. The most common frustration, cited by 66%, was AI answers that are "almost right, but not quite," and 45% said debugging AI-generated code takes more time.

What that means for junior developers

Read together, the numbers explain the pressure. AI is good at the tasks companies used to give juniors: boilerplate, simple features, tests and small fixes. It is unreliable in ways that take experience to spot. Companies are hiring fewer people to write first drafts, and still need people who can judge whether code is correct.

Skills that hold their value

Reading and reviewing code. A developer who can quickly find what's wrong with an "almost right" answer is useful from the first week.

Debugging and testing. Writing tests that prove behavior, and tracing a bug through a system, are where AI output most often needs a human.

Understanding the system. Knowing how the application's parts connect, how data flows and where security matters is context an AI tool doesn't have.

Working with the business. Turning a vague request into a clear specification is still mostly human work, and AI makes good specifications matter more, because they produce better generated code.

What employers should do

If juniors no longer learn by writing simple code, companies need another route to build senior engineers. Pairing juniors on code review, giving them ownership of tests and debugging, and keeping some work AI-free for learning are approaches teams are trying.