Why do entry-level jobs ask for years of experience?

People applying to those postings describe it as a closed door. One of them wrote: "How's someone supposed to get started if every beginner role wants you to already know the job?" The thread.

The replies on that thread are theories: employers can be picky, the line says preferred rather than required, companies would rather wait than train. Those are comments, not a measurement. We do not pick one and publish it as the cause.

A graduate asking whether to apply anyway is a different question. Career-office pages often say yes. One librarian's site suggests applying around 80 to 90 percent of the requirements. That percentage is their rule. It is not a finding, and it is not ours.

Three sources, in their words. The full lines and the links are on the proof library.

The International Labour Organization, Working Paper 140: one in four workers are in an occupation with some generative-AI exposure, and transformation of jobs is the most likely impact. Exposure is not replacement. (E1.)

The U.S. Bureau of Labor Statistics: high or very high relative AI exposure does not necessarily mean employment will decline. The same projections release says the U.S. economy is projected to add 5.9 million jobs from 2025 to 2035, growing 3.5 percent. That is the economy, not your odds. (E6.)

Stanford Digital Economy Lab, with ADP payroll records, revised August 2026: "We find no evidence of widespread, economy-wide job displacement." Employment of workers ages 22 to 25 in AI-exposed occupations "stands 19% below where it would be had it kept pace with that of their less-exposed peers." The authors call these facts descriptive indicators, not causal estimates. (E20.) Do not read 19% as jobs erased. Earlier drafts of the same paper said 13% and then 16%. Those figures are superseded.

A web search on 24 September 2026 for whether AI will replace entry-level jobs returned pages that repeat a Goldman estimate of hundreds of millions of jobs exposed. We have not read that paper at source. It is not used here.

Separate the posting's lines into ones your record already covers and ones it does not.

This person is fictional. The lines are an example.

The posting asks for a year of production support and for someone who has tested a user flow. Sam's record is a campus IT desk and a course project with a written test plan. The test plan speaks to the second line. The campus desk does not become a year of production support by changing the verb.

Apply or not is Sam's decision. The honest output is the split, not a percentage of qualifications that "should be enough."

Occupation tasks, as the public data lists them, are on the software engineer and data analyst guides. O*NET describes a task as a specific work activity that can be unique to the occupation. (E21.)

Paste the posting. Read which requirements your resume or your note already proves, and which it does not. Then practice saying the proof out loud, without a result the work did not have.

Job Match is 5 tokens. The question is free. A new account starts with 30 tokens. Neither one creates the year of experience the posting asked for.