Research check · 2 November 2026
One employer’s AI disclosure cut applications, not jobs
A research-seminar abstract hosted by HKU Business School reports that explicitly mentioning firm-wide AI adoption in job ads reduced applications by roughly one-third. The test covered 35 live vacancies and 3,865 applications at one industrial multinational. It did not measure a one-third fall in jobs.
What changed
The reported field experiment randomized whether vacancies disclosed the employer’s AI adoption. Applications fell most for coordination-intensive roles; direct-production roles showed little change. Candidates without observable technical skills were especially deterred. A companion survey experiment with 5,192 respondents found that people read the AI cue as possible role ambiguity and displacement risk. Adding an explicit offer of employer-provided AI training weakened the reported deterrent.
What the result can support
Randomized ad wording can test whether the message changed applications at this employer. It does not test whether the firm’s actual AI use changed vacancies, completed hires, layoffs or wages. Fewer applicants may change who enters a selection process without changing the number of jobs available.
What remains unverified
Only the institutional seminar abstract was accessible for this check. We could not inspect the full paper’s recruitment geography, treatment allocation, exact estimates, uncertainty intervals or hiring outcomes. The result should be treated as provisional until the paper is available and comparable employers are studied.