The risk in an automated no isn't the time it saves. It's what it tells the applicant.

You built the hiring pipeline last year, or you inherited it: job posting, applicant tracking system, and somewhere in the stack a tool that reads resumes and drafts the "thanks, but no" emails so nobody on your team has to write forty of them by hand. You approved it, or you never noticed it, and either way it has been sending your company's rejections since. You have probably never read one from the other side.

A new study looked at exactly that gap. Kyriaki Fousiani of the University of Groningen, with independent researcher Pieter A. Minnigh and Brian Lewis of Baruch College and the Graduate Center, CUNY, published "How rejection letters from AI recruiters shape recruiter's prestige and experienced respect" on September 7, 2026, in Frontiers in Organizational Psychology. It is not a study of real hiring. It is two online, vignette-based experiments run through the research platform Prolific: participants were told to imagine applying for a Project Coordinator role, to picture an AI recruiter reviewing their application and interview, and then to read the rejection letter it sent. Some got a "typified" letter — the bare minimum of thanks, decision, refusal, and sign-off. Others got a "considerate" version that added personalization and an explanation for the decision.

Study 1 ran in October 2025 with 197 UK adults, most of them currently employed, and manipulated whether they were told to imagine themselves unemployed. Study 2 ran in November 2025 with 232 adults, this time including 82 people who were genuinely out of work and job-hunting rather than only imagining it. In Study 1, the letter type alone did not move people's sense of having been processed like an object rather than considered as a person — what the researchers call "mechanistic dehumanization" — with any statistical reliability. In Study 2, the more realistic sample, it did: the flat form letter raised dehumanization scores significantly more than the considerate one. In both studies, once that feeling of dehumanization was present, it predicted a real drop in how much respect people felt they'd been shown and how much prestige they assigned the recruiter. But the full chain — letter type causing dehumanization causing lower respect and prestige — only held up statistically in Study 2. The first study didn't replicate it. That's worth sitting with before treating this as settled: it took the more realistic sample, with actual job seekers in it, for the effect to show up cleanly.

"The findings demonstrate that how AI recruiters communicate rejection can shape applicants' perceptions of the recruiter and their experience of respect."Fousiani, Minnigh, and Lewis, in the paper's Discussion, Frontiers in Organizational Psychology (read at frontiersin.org)

Here is why this lands differently for you than for anyone else in the building. An HR coordinator at a large company sends a template because it's the template; the decision about what it says was made by someone else, years ago, in a department they don't work in. You don't have that distance. At a company like yours, the rejection letter that goes out under your name is a decision you made, or a default you accepted, either way it's yours. Every one of those emails is doing reputation work you never see — with people who might become customers, might refer your next hire, might sit on the other side of a negotiation with you in five years. The study measured respect and prestige as abstractions on a survey. In your world those abstractions have names and LinkedIn profiles.

The part the research doesn't say, but that I'll say plainly: this study is about job candidates, not employees, and I'm not going to claim it proves anything about performance reviews or termination letters or the templated reply you send an unhappy customer. But the underlying trade it isolates — a few seconds saved for the sender against how disposable the recipient is left feeling — is not unique to hiring. It shows up anywhere a growing company starts routing hard conversations through a template instead of a person. That's my inference, not the study's finding, and I want to keep the two separate.

Here's the concrete thing to do this week: pull the actual text of the rejection your ATS or AI tool sends, and read all of it, the way a rejected candidate would. If it's the typified version — thanks, decision, refusal, done — add one true sentence: what stood out about them, or specifically why the role went to someone else. It costs you thirty seconds per template, not per candidate, and the study's own numbers say that's the difference between a company that processed someone and one that considered them.

If you want a sharper picture of where these blind spots tend to cluster in how you run the business — not just in what your software sends on your behalf, but in the decisions you make yourself under pressure — the Leadership Profile at /wiring-map is built for exactly that.