Make talent quality your leading analytic with skills-based hiring solution.
Candidate trust in job postings has eroded to the point where treating a listing as a good faith, straightforward offer is now the exception rather than the default assumption. This is not one problem, it is several separate failures that have compounded on top of each other over a short period. Only 8 percent of job seekers currently call AI hiring decisions fair, compared to 70 percent of hiring managers who trust AI to make faster and better decisions, a trust gap wide enough that Greenhouse has described the current environment as an outright AI trust crisis.
Rebuilding candidate trust in job postings starts with naming each contributing failure specifically rather than treating distrust as a single, vague mood problem.
A meaningful share of the postings candidates see are ghost jobs, roles published without genuine intent to fill them, covered in more depth in our companion piece on ghost jobs in 2026.
Every candidate who invests time applying to a posting that was never real becomes measurably more skeptical of the very next posting they see, and that skepticism does not stay isolated to the original employer, it generalizes across the market.
Separately, and compounding the same distrust, most candidates encountering AI in the hiring process report that its use was not clearly disclosed to them beforehand, with roughly 70 percent saying AI was not clearly flagged before their most recent AI involved interview or screening step, and roughly 80 percent describing employer AI policies as vague, rare, or entirely absent.
Only about 18 percent of candidates say most employers have explicit, clear policies governing how AI is used in their hiring process. Undisclosed AI is a distinct trust failure from a ghost job, but it produces the identical outcome: a candidate who no longer takes the posting or the process at face value.
The third compounding factor is simple non response. In one striking case, a job seeker with a master’s degree applied to thousands of positions over six months without a single callback, an extreme example of a pattern that is now common enough to shape how candidates approach every application, expecting silence rather than an outcome by default.
This ties directly to the volume dynamics covered in our piece on the AI hiring arms race, where average applications per job have climbed so sharply that most candidates, even strong ones, simply never hear back.
None of these three causes are fixed by a generic transparency statement. Ghost jobs are fixed by closing stale requisitions and disclosing pipeline building postings honestly. AI disclosure is fixed by stating plainly, in the posting or at the start of the process, when and how AI is involved, and by giving candidates a clear path to a human reviewer when they ask for one, a preference roughly 46 percent of candidates report wanting. And application silence is addressed by moving to a process that gives every genuine candidate a real, fast signal on where they stand, which is precisely what a structured skill assessment platform is built to support, an actual result a candidate receives rather than an application disappearing into an unexplained queue.
Candidate trust in job postings did not erode from one single failure, it eroded from three separate, compounding ones, fake postings, undisclosed AI, and unexplained silence, that candidates now assume by default rather than treat as the exception.
Employers who fix each cause specifically, rather than issuing a general transparency promise, are the ones who will stand out in a market where skepticism has become the baseline.
Candidate trust has eroded from three compounding causes: ghost job postings published without real intent to fill them, AI used in screening or interviews without clear disclosure, and widespread silence after candidates apply, with only 8 percent of job seekers calling AI hiring decisions fair.
Major causes include ghost jobs, undisclosed AI screening or interviewing, and application silence, where candidates receive no outcome or response after applying, a pattern common enough that many candidates now expect it by default.
Trust is low. Only about 8 percent of job seekers currently call AI hiring decisions fair, compared to 70 percent of hiring managers who trust AI to make faster and better decisions, and roughly 80 percent of candidates describe employer AI policies as vague, rare, or absent.
Employers can close or clearly relabel stale job postings, disclose plainly when and how AI is used in the hiring process, offer a path to a human reviewer on request, and give every candidate a fast, real result rather than silence after applying.
Exact figures vary by market and role, but non response has become common enough that some candidates apply to thousands of positions over months without a single callback, a pattern strongly linked to the sharp rise in average applications per job.
Skillfishing describes candidates, and sometimes existing employees, who overstate their capabilities, presenting a version of themselves on paper that looks strong but does not hold up once the actual work begins. The term is new, but the underlying research behind it is not a fringe finding. A full 91 percent of HR professionals now believe […]
This Communications Assistant job description template can be posted to online job boards and career pages to attract suitable candidates. The Communications Assistant job description, responsibilities, and requirements provided here can be customized based on your company’s needs. Key Responsibilities Job Statement We are looking for a Communications Assistant to support our communication efforts across […]
This Chief Marketing Officer job description template can be posted to online job boards and career pages to attract experienced candidates. The Chief Marketing Officer job description, responsibilities, and requirements listed here can be tailored based on your company’s goals and industry. Key Responsibilities Job Statement We are looking for a strategic and experienced Chief […]
