Blog / Hiring Resources
Built on Teamed ResearchWhy AI Skills Shortage Statistics Can Be Misleading
A large apparent gap can disappear when candidate documents and job postings are compared over the same period. That doesn’t prove a skills shortage doesn’t exist.
Key points
- Compared across mismatched years, 10.3% of candidate documents and 45.8% of job posts mention AI.
- Matched to 2026, the gap disappears: 46.7% of candidate documents versus 42.8% of job posts.
- A mention is not proficiency, and a matched year is not a matched job. Treat shortage claims as hypotheses.
An eye-catching number is easy to produce: compare the share of AI mentions across a multi-year résumé archive with the share in today’s job advertisements.
In Teamed’s records, 10.3% of documents in the multi-year candidate archive mention AI, while 45.8% of job postings in a separate employer measure mention it. Put side by side, those figures appear to describe an enormous shortfall.
But the comparison mixes documents created across different years with more recent job-posting language. It measures two different time distributions, not a direct skills gap.
What changes when you match the year?
In the 2026 cuts reported by Teamed, 46.7% of candidate documents written in 2026 mention AI, and 42.8% of job postings published in 2026 mention AI.
Looks like a 35-point gap
The gap is no longer present
The dramatic mention-rate gap is no longer present. Candidates actually mention AI slightly more often in these year-matched samples.
That’s the important finding—but don’t overread it.
What the comparison doesn’t tell us
- A mention in a résumé is not proficiency
A mention of AI in a résumé is not evidence of proficiency.
- A mention in a job ad is not a required skill
A mention in a job ad is not necessarily a required skill.
- Matching the year is not matching the job
Year matching does not account for differences in occupation, seniority, geography, employer sector or whether qualified applicants are available for a particular job.
So these figures do not establish that supply matches demand, that the market has enough AI-capable professionals, or that hiring difficulties are caused by employer screening.
They establish something narrower and still useful: this particular comparison between an older document archive and current job-posting language is not a sound way to quantify a skills shortage.
How to interrogate a shortage claim
Before repeating an AI skills-gap statistic, ask:
- Are the candidate and employer records from comparable dates?
- Do they cover the same jobs, seniority levels and locations?
- Are both sides measuring the same thing—mentions, requirements, demonstrated competencies or hires?
- Are duplicates, boilerplate mentions and classification errors handled consistently?
If those answers aren’t available, treat the claimed shortage as a hypothesis, not a measurement.
Where the research should go next
The next useful comparison is job family by job family: which AI capabilities employers explicitly require, how candidates demonstrate those capabilities, and where placements or assessments show genuine bottlenecks.
Figures are from different source populations; the discussion above is about comparability of text measures, not a direct estimate of market-wide AI skill supply or demand.
Keep reading
All articles →
Why Most L&D Job Posts Mention AI Without Naming a Tool
Employers often describe the work they need done without specifying the AI product. That creates both flexibility and ambiguity in hiring.
How to Build an L&D Team: Which Roles to Hire First
Building a Learning & Development team from scratch presents a difficult hiring question. Who do you hire first when the organization needs several different types of…
Designing a Hiring Process That Measures Learning Impact
Most hiring processes for learning and development roles still reward the signals that are easiest to see: strong resumes, familiar tools, polished portfolios, and confident interviews.…