Young adults aged 22–25 are bearing more pronounced impacts in roles with high AI exposure. Maria Black, President and CEO of US human resources and payroll services firm ADP, put forward this view in a recent discussion on AI and employment.

Young adults aged 22–25 are bearing more pronounced impacts in roles with high AI exposure.

Maria Black, President and CEO of US human resources and payroll services firm ADP, put forward this view in a recent discussion on AI and employment.

ADP processes payroll for roughly one-sixth of the US workforce and serves more than 1.1 million enterprises globally. Instead of abstract predictions over “whether AI will take jobs”, she observes real shifts in wages and employment for millions of American workers.

Her assessment does not claim “AI will not affect employment”. From an overall market perspective, AI has not triggered widespread job elimination. But broken down by age and occupation, young people are feeling pressure first-hand: if jobs are not disappearing across the board, why is it growing harder for new graduates to enter the workforce?

01 AI Is Not Erasing Jobs — It Is Unbundling Work

Maria Black describes the ongoing shift as the “unbundling” of work. Rather than wiping out entire professions outright, AI redistributes tasks within job roles. Some work is automated, some remains human-led, and certain responsibilities previously reserved for senior staff are now pushed earlier onto new hires.

Take entry-level analyst roles as an example. In the past, new recruits started with information gathering, data organisation and report drafting, learning the business through hands-on execution. Today, AI can complete portions of this work rapidly. Companies may not eliminate the “Analyst” job title entirely, yet they may cut hiring volumes or require new hires to validate outputs, interpret findings and drive implementation from day one. Similar changes are unfolding in software development, marketing and customer service.

The core shift: companies once hired juniors to **complete foundational tasks**. Now employers prefer to pay for people who can **define problems, judge outputs and take accountability**.

Entry-level roles have not vanished entirely, but the traditional on-ramp into the workforce is narrowing.

02 Why Young Workers Are Hit First

Updated research from the Stanford Digital Economy Lab, built on ADP payroll records, shows that among occupations with high AI exposure, employment for workers aged 22–25 sits roughly 19% below the level they would have reached had they grown at the same rate as peers in low-AI-exposure fields. Senior staff within those same professions have not experienced a comparable gap.

The study covers millions of US workers, with data updated through June 2026. The research team cautions these are early descriptive signals, not definitive causal proof that AI drives employment changes. Slower hiring, industry cycles and corporate cost controls also shape opportunities for young talent.

Basic coding, literature research, data cleaning and routine support queries have long served as the training ground for junior employees to learn business fundamentals. As AI takes over these tasks, companies still need people who understand clients, assess risks and solve complex problems — yet fewer entry-level stepping stones remain for new talent to build those capabilities.

This creates a new paradox: companies expect new graduates to handle more complex work immediately, while the junior roles that would let them build that complex experience are contracting.

03 US Employers Are Raising the Bar for New Grads

Many students’ first response to the AI trend is adding ChatGPT, Claude or Copilot to the skills section of their resume. Maria Black, however, notes that AI literacy is becoming a baseline productivity requirement rather than a lasting differentiator.

Simply knowing how to launch a tool does not prove you can deliver value. Employers care more about whether you can judge the reliability of AI-generated answers, spot risks in data, logic and compliance, and turn rough drafts into actionable solutions.

This means the competitive edge for future new grads will increasingly hinge on three core strengths.

1. **Domain judgement.** Finance candidates must evaluate whether model assumptions hold; marketing candidates assess audiences, channels and conversion; software engineers handle testing, security and real-world edge cases.

2. **Output validation.** In interviews, candidates must explain which datasets they used, how they identified errors, why they selected a given approach, and how outcomes are measured.

3. **Communication & stakeholder alignment.** As AI takes over some execution work, client understanding and cross-team collaboration grow even more critical.

ADP research also found that among employees using AI daily, 30% report high work engagement. For those who never use AI, that figure stands at only 14%.

04 How International Students Can Adapt to This Shift

Adapting does not mean everyone must switch to AI-related majors. A more practical path is rethinking how your own discipline intersects with AI.

Students in software development, data and AI tracks should demonstrate tool selection, data handling, model evaluation and audit procedures when outputs are unreliable.

Finance, consulting and business analytics students need to practise the full workflow from raw data to commercial conclusions.

Marketing, operations and customer success candidates should focus on user insights, experiment design, client communication and project delivery.

For majors such as EE, ME and supply chain, AI opportunities are not limited purely to software roles. Jensen Huang, CEO of NVIDIA, recently remarked that chip fabrication, packaging, computer manufacturing and AI data centre construction will create large numbers of engineering and technical positions.

Especially for the Class of 2027, here are key priorities to prepare for right now:

1. Revisit project descriptions on your resume. Projects should not merely prove you used Python, SQL or a large model; explain where the problem originated, why you chose your methodology, and how you validated results.

2. Gain hands-on experience in real-world contexts. Internships, research and live client projects expose students to constraints and ambiguous scenarios absent from classroom work.

3. Do not list AI as an isolated item in your tool checklist. Clearly document which workflows you improved, what human judgement you applied, and what measurable outcomes were achieved.

4. Practise reasoning and articulation in advance: Why this solution? What will you do if AI produces incorrect output? How do you push projects forward when team members disagree?

Maria Black’s remarks do not tell young people “AI will not affect you”. Instead, they reveal a subtler transformation. The impact may not appear as the disappearance of an entire profession, but as compressed foundational tasks, fewer junior openings and earlier performance expectations for new hires. Capabilities once built gradually after onboarding now increasingly need to be demonstrated at the recruitment stage.

For international students, the real concern is not whether “AI” appears on your resume. It is whether, when AI can handle large volumes of baseline work, you can still demonstrate domain mastery, independent judgement and the ability to solve real-world problems.

AI has not eliminated all jobs, yet it is redefining who gets that first professional role.