Everyone looks good on paper now

A job application once held weight: a well formed resume, a clear cover letter, a thoughtful screening answer. In the past two years, genAI has flattened most of that trusted signal to noise. Now, trust is eroded and talent professionals are struggling to make heads or tails of this new hiring landscape. The problem is not only resting on candidates, who are doing the logical thing by using AI tools to polish and mass apply to jobs. Employers are still using screening methods that were built for a hiring landscape that no longer exists.

This is not a candidate problem

The seductive response is to blame candidates, to say their use of AI tools to mass produce cover letters and applications has turned the process into an untrustworthy mess. But let’s be frank: people are doing the rational thing. According to Greenhouse’s 2026 AI in Hiring Report, 74% of candidates in the U.S. now use AI in their job search. They are using tools that make them shine, make applying faster and more targeted in a bid to increase their chances of being seen. Most of us, put in the same position, would do exactly the same.

The problem lies with employers, not applicants. By attempting to add AI or some form of automation to every stage of the hiring funnel, we’ve been the architects of our own demise, collapsing the very signals we once used as markers and reducing our ability to trust who and how we hire.

We are at a point where talent professionals must take a sobering look at how we are implementing new tools, and how we can salvage the current situation to build confidence earlier in the screening process. The screening stage is where serious hiring missteps start, everything else flows downwards from there. It’s incumbent upon us to make this change: we as employers own the system, so we must own the fix.

Catch more HRTech Insights: HRTech Interview with Emma Lavelle, Chief Operating Officer, UneeQ

What’s actually changed

Candidate behavior has been shifting for a decade, driven by macroeconomic pressure as much as technology. AI just compressed that shift into two years. The response from talent professionals has been reactive, the solution being adding more AI or tools to the mix.

Volume is the most obvious change: AI models can tailor an application to any job description in seconds. Candidates can apply to far more roles than they could before, meaning pipelines balloon. A recruiter opening a requisition on a Monday can have hundreds, if not thousands, of applications in 24 hours and be no closer to knowing who should progress. Sifting through these applications has become an onerous task, but the bigger problem is the lack of evidence and strong signals early on.

These tools have also homogenized the process. Since each resume has a high level of polish now, almost everything reads well and written communication has stopped being a useful filter for capability. Talent professionals are reading the model, not the person.

Then there are take-home tasks and assignments, which used to be one of the more legitimate ways to tell whether someone could actually do the work or not. AI works best in real time. It provides live assistance, so those take-homes, written exercises and even interview answers we relied upon are now rendered ineffective as these can all be completed with AI. If AI can do the test for you, you are testing the tool, not the person.

Calling it what it is: a screening crisis

I see this as a screening crisis. Most screening leans on polish and what looks good on paper as proxies for judgment of the hire. AI has made those proxies ubiquitous, and they no longer aid in separating anyone from anyone.

A controversial take, maybe: AI did not create this problem. It exposed how weak paper always was. A resume was never strong evidence of ability. It was the cheapest and most widely accepted evidence available, and we built entire hiring processes on top of it because nothing better existed at scale.

Screen for thinking, not output

If we are considering updating our systems to be more trustworthy and effective, we must stop screening for output and start screening for thinking. Reasoning and judgment are the trust signals now, so the process must encourage it over output. Create a structure that allows candidates, live or on video, to think through a decision, react to something they couldn’t have prepared in advance or even defend a trade-off. Direct assessment and hands-on validation are key ways to reduce the chances of the work being offloaded to a model.

Structure your processes to keep consistent, role-relevant questions for every candidate, reviewed against the same criteria, so you compare like for like and the final decision stays with a human.

The answer isn’t banning AI either. It’s here to stay, so make it part of the test: ask candidates to use it and show you how. Watching someone prompt, question and iterate tells you far more than pretending the technology is not in the room.

I’m not suggesting that we abandon any of these tools. Instead, our industry must become clearer about what they are for. AI is genuinely useful for making candidate evidence easier to gather and review at scale, but it’s a poor substitute for the judgment call itself, and every time we have let it stand in for that call, we have paid for it in trust.

The companies that hire well over the next few years will not be the ones with the best detection software, or the ones that have automated the most. They will be the ones honest about what a resume was ever really worth, and willing to rebuild screening around the human signal that is left.

About Willo

Willo is a Glasgow-based virtual interviewing and candidate screening platform launched in 2018

Read More on Hrtech : Agentic HR: Can AI Become a Workforce Strategist Instead of Just an Automation Tool?

[To share your insights with us, please write to psen@itechseries.com ]

The post Everyone looks good on paper now appeared first on TecHR.