Why job descriptions now read like CVS receipts

When Robin Olsen started applying for communications jobs three months ago, she noticed a shift. The job descriptions had grown longer. One combined communications, marketing, and public relations into a single block of text that read more like a wish list than a real job offer. "It felt like the company didn't know what they wanted," she said.

Olsen is not alone. Across industries, job seekers report that postings have ballooned in length and scope. The culprit, recruiters say, is the same tool many companies are using to write those descriptions: generative AI.

Large language models excel at producing dense, comprehensive lists. When a hiring manager asks an AI to draft a job description, the model pulls from similar roles across the internet. It often merges requirements from senior and junior positions, adds nice-to-haves that were never core, and lists every tool or platform that appears in any comparable listing. The result is a document that can run five hundred words or more, with a bulleted section of qualifications that seems to cover everything.

That creates a problem for applicants. A long list of requirements discourages qualified candidates. Studies have shown that women, in particular, are less likely to apply unless they meet nearly every criterion. A description with 15 required skills can shrink the applicant pool before a single resume is read.

Employers face the opposite issue. More applications pour in, the match rate drops. Candidates interpret the description loosely, hoping some of the skills stick. Recruiters then must spend extra time screening, which in turn leads them to adopt their own AI tools to parse resumes. A feedback loop develops: AI writes a bloated JD, AI screens the responses, and the human touch gets squeezed out.

Some companies have started pushing back. A handful of hiring managers now explicitly tell job seekers to ignore the full list and apply if they meet half the requirements. Others require the AI-generated draft to be edited down to a hard cap of 10 bullets. Those practices are not widespread.

For job hunters like Olsen, the best strategy is to read past the first paragraphs. "Look for the verbs that keep coming up," she said. "If the same core task appears three times in different phrasing, that's probably the real job."

The trend is not likely to reverse soon. More companies are integrating AI into their HR workflows, and the tools only get better at generating long, plausible descriptions. Whether the market adapts with shorter, human-edited postings – or simply learns to ignore the extra noise – is still an open question.

Olsen eventually landed an interview. The job description had been short.