Two surveys of recruiters landed this week, and together they say something every leader should hear about AI in hiring: the tools made it faster to find people, but they did not make it easier to know who those people are. Speed went up. Confidence did not.

That is not a technology failure. It is a signal failure. A polished resume used to tell you something. Now it mostly tells you that someone had access to a tool.

And it is not only a hiring story. A new leadership study released the same week suggests the same gap sits inside our organizations, not just at the front door.

Key takeaways

  • AI in hiring has made sourcing candidates faster, but recruiters report that their confidence in candidates has not kept pace.
  • AI-polished resumes are now the top stressor for recruiters in Employ’s long-running survey, ahead of a lack of qualified candidates.
  • Polish used to be a signal of effort and care. AI has made that signal cheap, so hiring needs explicit standards and visible judgment instead.
  • The same gap shows up among leaders: most report AI productivity gains, but few say leaders check the quality of AI output.

What’s happening

On October 16, HR Dive senior editor Kathryn Moody reported data that the recruiting platform TestGorilla shared from a survey of more than 1,400 U.S. talent professionals (recruiters find people faster with AI, but confidence in candidates hasn’t followed). Seventy-two percent said AI made finding candidates quicker. But about a third of those respondents said they gained no more confidence in candidates’ qualifications, and 18% said their confidence fell. More than two in three said they spend an hour or more manually checking a candidate’s credentials, and 30% said they spend two hours or more.

TestGorilla’s CEO, Wouter Durville, put it in one line: “Recruiters got a better search engine and a worse index in the same 12 months.” Since that data comes from a hiring-tech vendor, I read it as one useful window rather than the final word. But it does not stand alone.

The same day, HR Dive’s Lara Ewen covered Employ’s survey of more than 1,200 U.S. recruiters and hiring managers, now in its 19th year (AI is the root of recruiters’ biggest stressors). For the first time in the study’s history, AI is the root of recruiters’ biggest pain points. One-third named AI-polished resumes as their main source of stress, which the report says causes more headaches than a lack of qualified candidates. Meanwhile, 68% use AI alongside their other hiring software and 66% said it made their jobs easier, yet only 18% said more than half of their time has shifted from administrative work to higher-value tasks. Employ CEO Jerry Jao said teams can now assess whether their processes, technology and AI investments are working, “Or if they’re working at all.”

Then there is the view from inside the organization. On October 15, HR Dive reported on DDI’s Global Leadership Forecast 2027, which draws on more than 11,000 leaders and about 2,500 HR professionals (leaders say they are worried about their ability to handle change). Eighty-one percent of leaders cited AI productivity gains and 75% reported better work quality. Yet only 14% felt ready to use emerging technologies successfully, and only 18% said leaders check the quality of AI output.

Put those together and a pattern appears. AI is producing more, faster, and more polished. The checking has not kept up.

What AI in hiring reveals about broken signals

In The Next Turn, I describe two changes happening at once. The definition of work is changing: what people believe work is for and what it should provide. The structure of work is changing too: how work actually gets done now that AI is part of the infrastructure. This week’s news sits squarely on the structural side. It is about how work flows, and what that does to the evidence we rely on.

In Chapter 7, “The Collapse of Effort,” I argue that effort has long served as a proof signal. Hours, drafts, visible struggle. We read them as evidence of commitment and quality. AI breaks the relationship between visible effort and visible value. Hiring has its own version of that signal. A carefully written resume and a tailored cover letter used to suggest time, care and communication skill. Now a candidate can produce both in minutes.

Not the collapse of work. The collapse of effort as a signal.

The Next Turn

Notice what that does not mean. It does not mean candidates are less capable or less honest. Using AI to write a resume is a rational response to a market where, as HR Dive notes, Ashby found applications per hire have roughly tripled since 2021. Candidates of every age are adapting to the structure they are in. The resume did not get worse. It got less informative.

Chapter 11, “Where It Breaks,” names hiring as one of four pressure points where the double transition becomes operational. When the old proof system thins out, leaders tend to reward what is easiest to see. In hiring, what is easiest to see is polish. And polish is now abundant.

This is where Chapter 8, “The Rise of Judgment,” matters. Polished AI output can create the illusion that judgment has been exercised. That is true of a candidate’s resume, and it is just as true of the recruiter’s AI-generated shortlist or a leader’s AI-drafted analysis. DDI’s finding that only 18% say leaders check AI output quality tells me the hiring problem is a mirror. We are asking whether candidates used judgment while we are not always showing ours.

What increasingly separates people is not whether they can produce a first version. It is whether they can decide what that version is worth.

The Next Turn

Durville’s other line in the TestGorilla coverage gets at the same thing: “The expensive thing is a shortlist you can defend to a hiring manager.” Defend is the right word. A defensible decision is one where the reasoning is visible.

Standards, not signals, in the hiring process

In Chapter 13, I separate standards from signals. Standards are what the organization actually needs: quality, reliability, sound judgment, follow-through. Signals are the visible cues we use to infer them. A polished resume was always a signal. So was a confident interview, a fast reply, a long list of keywords.

The standards remain. The proof system changes.

Recruiters spending an hour or two per candidate checking credentials by hand are already living this. They are rebuilding the proof system manually because the old one no longer holds. The leadership question is whether we keep patching that gap with more hours, or redesign the process so it tests for the standards directly.

Your next turn

  • Name the standard behind each screen. For one open role, list what each step of your process is actually trying to learn. If the answer is “writing quality” or “attention to detail,” ask whether a resume can still show that.
  • Move evidence closer to the work. Replace one signal-based screen with a short, realistic task and a conversation about it. Ask the candidate to walk you through the choices they made, and listen for reasoning, not polish.
  • Define AI’s role in your hiring, on both sides. Decide what you expect from candidates who use AI, and say it in the posting. Then decide where your own team’s AI tools help, and where a human must review before anyone is screened out.
  • Check the checkers. If your team uses AI to summarize, rank or shortlist, pick five recent decisions and have someone trace the reasoning. Could you defend each shortlist to a hiring manager?
  • Ask where saved time went. If AI saved your recruiters hours, decide on purpose whether those hours go to deeper conversations with fewer candidates, rather than more volume.

Frequently asked questions

Should we reject candidates who use AI to write their resumes?

Probably not. Recruiters themselves report using AI widely, and candidates are responding to a crowded, high-volume market. A better move is to be clear about what you expect, and to design later steps of the process that show the candidate’s own thinking.

How do we evaluate candidates when AI makes every application look polished?

Shift from signals to standards. Decide what the role actually requires, then test it closer to the real work, through a short task, a structured conversation about decisions, or a work sample discussed live. Ask how they decided, what they verified, and what they would do differently.

Is AI in hiring saving time or creating more work?

Both, according to this week’s surveys. TestGorilla’s data showed 72% found sourcing quicker, but many recruiters spend an hour or more verifying each candidate by hand. Employ found that only 18% said more than half their time had shifted to higher-value work. The time savings are real, but they do not automatically become better decisions.

The real test

AI did not lower the bar for candidates. It removed one of the shortcuts we used to tell whether someone cleared it. That is uncomfortable, but it is also clarifying. It forces us to say out loud what we are actually looking for.

The same is true inside our teams. If only a small share of leaders are checking AI output, the most important hiring standard might be the one we model ourselves.

So here is my question for you: if every resume on your desk looked equally polished tomorrow, how would you know who was ready for the work?

Sources