AI skills for entry-level jobs: what new graduates should show

New graduates should show that they can take an entry-level task further than an AI tool can on its own. That means checking the output, combining steps and delivering finished work. Hiring of young workers has fallen in jobs where AI can do the tasks alone. It has held up better where AI helps people work. Your aim is to show you belong in the second group.

What the data says about entry-level hiring

Researchers at the Stanford Digital Economy Lab studied ADP payroll data for millions of US workers (Stanford). It looked at workers aged 22 to 25 in the most AI-exposed jobs. By mid-2026, their employment was 19% below where it would have been had it kept pace with less-exposed jobs. Experienced workers in the same jobs showed no such gap. The gap came mainly from fewer young people being hired, not from more people leaving.

The same study split jobs by how AI is used in them. Declines were concentrated where AI mostly replaces human tasks. Where AI mostly helps people do their work, employment was flat or rising. The authors describe these as early signals, not proof of cause.

Indeed Hiring Lab saw a related pattern in software jobs. US software development postings rose almost 15% from February 2025 to June 2026 (Indeed Hiring Lab). Senior roles made up 71% of the rise between May 2025 and May 2026.

What the data does not tell you

• Both studies use US data. Entry-level hiring in other countries may follow a different path.
• AI exposure varies within a job title. Two graduate roles with the same name can involve different tasks.
• The payroll study shows employment levels, not why a given employer hired fewer juniors.
• Other forces also affect graduate hiring. The Stanford authors call their results descriptive, not causal.

The practical point holds even so. Tasks that AI can do alone are a weak basis for an entry-level application. Tasks where you add a check or a decision are a stronger one.

What employers say they want

The 2024 Work Trend Index from Microsoft and LinkedIn surveyed 31,000 people in 31 markets (Microsoft). In it, 71% of leaders said they would rather hire a less experienced candidate with AI skills than a more experienced one without them. And 77% said early-career staff will be given greater responsibilities because they can hand more work to AI.

Read together, the two sets of findings point in one direction. Employers may hire fewer juniors for tasks AI can do alone. They still want juniors who can use AI to take on bigger work sooner. The skills that matter are the ones that turn a draft into a result.

Five skills to show

1. Checking output. Find the wrong figure, the invented source or the missing step before anyone else sees it. A tool that only drafts cannot do this for itself.
2. Giving context. Supply the brief, the data and the limits. A junior who gathers the right material first gets better output.
3. Combining steps. Break a task into research, draft and check, using AI where it helps. Write down the steps so someone else can repeat them.
4. Finishing the work. Deliver something a manager can use without rework: a formatted summary, a cleaned dataset, a reply ready to send.
5. Keeping data safe. Know what you may paste into which tool. Ask about the policy on your first day.

Each skill has a free study guide: checking the work, giving context, putting it together, getting work done and keeping it safe.

Entry-level tasks to practice on

Practice on the tasks that junior staff do in your target field. Use public data or invented data. Some examples:

• Marketing: draft three versions of a product email, then check every claim against the product page.
• Support: draft replies to five customer questions, then check each one against the help center.
• Operations: turn a messy meeting transcript into a list of actions with owners and dates.
• Data: clean a public dataset with AI help, then spot-check 20 rows by hand.
• Sales: research a public company and write a one-page account brief, with a source for each fact.
• Legal or compliance: summarize a public regulation, then check each point against the text.

The role guides list more AI tasks for eight functions.

Make the most of internships, projects and coursework

An internship gives you real tasks, which are the best evidence. Use it on purpose.

• Ask your manager in week one which AI tools are approved and what data they may see.
• Keep a private log of tasks you did with AI, what you checked and what you changed. Leave out confidential details.
• Ask whether you may describe one task in general terms in future applications.
• If AI is not allowed at the internship, note the manual tasks you would now speed up with AI. That reasoning is still useful in an interview.

Without an internship, a class project or a volunteer task works too. Follow your institution's rules on AI use in coursework. If a course bans AI, do a separate practice task instead.

How to show it to an employer

• On the resume, describe a task and a check, not a tool. "Summarized 40 survey responses with AI and verified each theme against the raw data" says more than "ChatGPT".
• Bring two or three work samples to interviews. See how to build an AI work portfolio.
• In an interview, tell one story about an AI error you caught. Judgment is hard to see in the output alone, so a story shows it.

For wording, see AI skills to put on your resume. For what the term AI literacy means to employers, see what AI literacy means for job seekers.

An independent result can add a second view to your own account. Emeri offers a free 15-minute AI skills check. Emeri certification is assessed on real work tasks and reported as a band from E1 to E4 (bands).