→Claude Just Joined Your Meetings, But Only on Google Meet →An AI Vendor Breach Just Leaked Uber Driver IDs and Notion Chat Logs at the Same Time →A Safety Reminder Cut Harmful AI Medical Advice by a Third. It Still Got 1 in 10 Wrong. →A Power Bank Just Launched That Costs 75 Percent More to Not Catch Fire →Firefox's New Mini Window Is Genuinely Useful, With One Annoying Catch →How to Actually Check if Your Email or Password Was Leaked in a Data Breach →Google's New AI Model Reads DNA Cell by Cell, and the Numbers Are Real →Perplexity's Comet Browser Is Free Now: What You Gain and What You're Trusting It With →Dropbox Just Spent $350 Million on a Storage Startup to Escape Being Dropbox →AI Is Not Taking Your Job. It Might Just Be Freezing Your Raise →NotebookLM Is Gone, Gemini Notebook Is Here: What Actually Changed →Four AI Models, Five Accents, One Hierarchy: What a New Study Found →Claude Just Joined Your Meetings, But Only on Google Meet →An AI Vendor Breach Just Leaked Uber Driver IDs and Notion Chat Logs at the Same Time →A Safety Reminder Cut Harmful AI Medical Advice by a Third. It Still Got 1 in 10 Wrong. →A Power Bank Just Launched That Costs 75 Percent More to Not Catch Fire →Firefox's New Mini Window Is Genuinely Useful, With One Annoying Catch →How to Actually Check if Your Email or Password Was Leaked in a Data Breach →Google's New AI Model Reads DNA Cell by Cell, and the Numbers Are Real →Perplexity's Comet Browser Is Free Now: What You Gain and What You're Trusting It With →Dropbox Just Spent $350 Million on a Storage Startup to Escape Being Dropbox →AI Is Not Taking Your Job. It Might Just Be Freezing Your Raise →NotebookLM Is Gone, Gemini Notebook Is Here: What Actually Changed →Four AI Models, Five Accents, One Hierarchy: What a New Study Found
Panda Tech Bytes
All Posts
Latest AI Trends Guides Tools Tech Pulse Hardware & Gear Funding Watch Risk & Reality What It Means for You Security Watch On Your Desk
Stanford Tracked Entry-Level Hiring for Four Years. Here is What Changed

Stanford Tracked Entry-Level Hiring for Four Years. Here is What Changed

A Stanford study using real ADP payroll data finds the employment gap for young workers in AI-exposed jobs has grown from 13% to 19% in a year, but the researchers themselves warn against reading too much into that number.

If you are 22 to 25 years old and applying for your first job in software, customer support, or other entry-level office work, there is now a real, data-backed reason hiring feels harder than it used to. Researchers at Stanford's Digital Economy Lab have been tracking payroll records since ChatGPT launched, and their latest update, released in August 2026, found that employment among young workers in the most AI-exposed occupations sits about 19% below where it would be if it had kept pace with similarly aged workers in less-exposed fields. A year earlier, that same gap was 13%. It is widening, and the researchers can tell you why it is probably happening, even if they are careful not to overstate what they have found.

What the study actually measured

The paper, titled "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence," comes from Stanford economists Erik Brynjolfsson, Bharat Chandar, and Ruyu Chen. They used anonymized payroll data from ADP, the largest payroll processor in the United States, which gives them a far larger and more frequent dataset than typical government employment surveys.

They sorted occupations by how exposed they are to generative AI tools, using existing task-based exposure scores, then compared employment trends across age groups within each exposure level. The headline comparison is simple: workers aged 22 to 25 in highly AI-exposed roles versus workers the same age in roles AI barely touches.

The numbers behind the gap

Between November 2022 and roughly mid-2026, employment for young workers in the most AI-exposed occupations fell by around 11%, while employment for the same age group in less-exposed occupations grew by about 10%. Software development is one of the clearest examples: employment for young software developers was down nearly 20% from its late 2022 peak by mid-2025, even as demand for experienced developers held up.

Customer service and clerical roles show a similar pattern. The common thread across all of them is what the researchers call "codified knowledge," meaning tasks built on formal, documented procedures that can be taught from a manual or textbook. Those are exactly the tasks generative AI is best at imitating.

Notably, experienced workers in the same occupations show no comparable gap. The adjustment appears to be happening almost entirely through reduced hiring of young workers, not through layoffs of people already in the job.

Why this is not the same as "AI is destroying jobs"

The researchers are unusually blunt about what their data does not show. They explicitly state there is "no evidence of widespread, economy-wide job displacement" from AI. The effect they are describing is narrow: a specific age group, in a specific set of occupations, losing ground on hiring relative to their peers. It is not a broad labor market collapse, and the authors describe their findings as "early, descriptive indicators, canaries in the coal mine, rather than causal estimates."

That distinction matters. A company not posting as many junior roles because AI tools now handle first-draft code review or basic support tickets is a very different story than a company laying off its existing junior staff. Most of what this study captures is the former.

The caveats worth knowing before you repeat this stat

Independent reviewers who dug into the methodology found several things worth flagging. The gap shrinks noticeably once you control for education levels, meaning some of the hiring slowdown may reflect broader credential and degree trends rather than AI specifically. Some of the decline also predates the release of generative AI tools, which complicates a clean before-and-after story.

The size of the effect is also larger in ADP's payroll records than in national household surveys like the Current Population Survey, which showed a much smaller decline over the same period. That gap between data sources suggests the true size of the effect is still genuinely uncertain, even if the direction (young workers in AI-exposed jobs losing ground) shows up consistently.

What this actually means if you are job hunting or hiring

If you are early in your career and targeting a role built around repetitive, well-documented tasks, standing out now probably means being able to show you can do the parts of the job that are not easily codified: judgment calls, communicating with frustrated customers, debugging something that is not in the manual. Those are the "tacit knowledge" tasks the study found AI still struggles to replace, and they are exactly where experienced workers in the same study showed no employment gap at all.

If you are on the hiring side, this is also a flag about pipeline risk. A company that stops hiring junior employees because AI tools cover the easy 80% of the job is quietly skipping the step where junior employees become senior ones. That is a slower-moving problem than a layoff, but it is a real one for any business that will eventually need experienced staff five years from now.

The takeaway

The honest summary of this research is narrower than the headlines about it: young workers in a specific slice of AI-exposed jobs are getting hired less, not fired more, and the researchers themselves are not calling it proof that AI is reshaping the whole labor market. What makes this study worth paying attention to is not a dramatic single number, it is that the same careful team has now tracked this gap widening for two straight years using real payroll data, while also being transparent about exactly where their evidence gets shakier. That combination, a real trend plus honest uncertainty about its size, is rarer than either a panic headline or a dismissal would suggest.