The Door That's Closing First: Why AI Is Hitting Entry-Level Jobs

If you are about to graduate or have just graduated, this is probably the most personal post in the series. The evidence suggests that AI is affecting jobs unevenly, and that people at the start of their careers feel it first. Here is what the research says, and what you can do about it.

What the Stanford study found

The Stanford Digital Economy Lab published an August 2026 update to its "Canaries in the Coal Mine" research, using ADP payroll data. It compared the employment of 22 to 25 year-olds in highly AI-exposed jobs against peers in less-exposed jobs. In July 2025 the gap was 15%. By June 2026 it had widened to 19%. For experienced workers, there was no comparable gap.

The headline is careful: no widespread displacement, but the AI employment gap for young workers has widened. That is an honest summary. The sky is not falling for everyone, but the first rung of the ladder is getting harder to reach.

How it happens

It does not look like mass firing. Companies simply hire fewer juniors. Where AI automates tasks, those jobs shrink. Where AI augments people, employment is flat or rising.

Why juniors first

Juniors are hired for codified knowledge: the things in textbooks and courses. That is exactly what AI is good at. Tacit knowledge, such as judgment, experience and knowing how an organisation works, is gaining value. An experienced analyst knows which number looks wrong. That is hard to learn from a course.

The graduate picture

The New York Fed reported that computer engineering graduates in the US have an unemployment rate of 7.8%, the second highest among college majors, as VnExpress reported. The World Economic Forum published a briefing on AI and entry-level jobs in January 2026 that looks at the same issue.

These are US numbers, and the Pakistani market differs. But many Pakistani graduates compete for remote and multinational work, so the pressure travels.

What a graduate can do

1. Show tacit skill early. Internships, freelance work and open-source contributions prove you can work with real constraints.
2. Be a visible AI user. Document how you used AI to deliver something faster, and what you checked.
3. Pick roles where AI augments. Data, security, product and operations roles with human judgment are safer than pure routine production.
4. Build a small portfolio of three projects you can explain, not ten that you cannot.
5. Get real data. Work with messy, real-world data sets, not only tidy classroom ones. Cleaning, questioning and explaining data is a skill employers still struggle to find.

If you want a structured, zero-cost way to start building those skills, read the free AI learning roadmap. For the broader picture of how tasks and roles are changing, see Shrinking Tasks Versus Growing Roles.

Keep reading

• "AI" Is Now the #1 Reason Given for US Job Cuts: Reading the Layoff Numbers Honestly
• Shrinking Tasks vs Growing Roles: Which Jobs Are Under Pressure and Which Are Growing
• World-Class AI Education for Rs. 0: A Free Four-Step Learning Roadmap

See the complete AI Era series

---