Each recruiter who has opened their ATS prior to now yr has felt it: one thing has shifted. The functions look totally different. They learn smoother, hit extra key phrases, and sound eerily much like each other. That uniformity just isn’t a coincidence. In line with a 2025 report from Profession Group Firms, roughly 65 p.c of job candidates now use AI in some unspecified time in the future throughout the utility course of. Resume Now’s survey of over 900 HR professionals discovered that 90 p.c have seen a rise in low-effort or formulaic functions, largely pushed by AI instruments.
The result’s a real operational downside. Recruiters are managing 2.7 instances extra inbound functions than they have been three years in the past, whereas crew headcounts have truly shrunk. The previous screening playbook — scan for key phrases, verify formatting, search for a compelling cowl letter — was constructed for a world the place these alerts nonetheless meant one thing. That world is disappearing quick.
So what can we do about it?
The Scale of the Shift
Let’s be trustworthy concerning the numbers. Utility volumes per job opening have surged roughly 182 p.c since 2021, in accordance with recruiting benchmarks knowledge from Ashby. A few of that enhance is pushed by “simple apply” options and a aggressive labor market, however a rising share is attributable to candidates utilizing generative AI to supply and submit supplies at scale.
This creates a paradox. AI instruments on the employer aspect have been supposed to assist handle quantity, and in some methods they do — automated screening can scale back resume evaluate time by as much as 75 p.c. However when the functions themselves are additionally AI-generated, these screening instruments are more and more matching on surface-level polish somewhat than substance. You find yourself with a system the place AI is screening AI, and the people within the loop are left attempting to determine who truly wrote what, and whether or not it issues.
For recruiting groups already stretched skinny, this isn’t an summary concern. It’s a workflow downside that instantly impacts high quality of rent.
Why “Ban AI” Is Not a Viable Technique
Some organizations have responded by asking candidates to certify that their supplies are “AI-free.” This method is comprehensible however in the end unworkable, for 3 causes.
First, enforcement is sort of unimaginable. Detection instruments stay unreliable, producing frequent false positives that may penalize robust writers and non-native English audio system who occur to put in writing in a method that triggers the algorithm. Constructing a screening course of on a basis of inaccurate detection is a compliance danger, not an answer.
Second, the road between “AI-generated” and “AI-assisted” is already blurred past recognition. A candidate who makes use of Grammarly, a spell checker, or a sensible template is utilizing AI. A candidate who asks ChatGPT to brainstorm learn how to body a profession transition, then writes the precise content material themselves, is utilizing AI. Drawing a vivid line between acceptable and unacceptable use is a shedding proposition.
Third, and most significantly, banning AI use in functions sends an odd sign in 2025. Most organizations are actively investing in AI adoption throughout their operations. Telling candidates they can not use the identical instruments your personal workers depend on each day creates a credibility hole that savvy candidates will discover.
The higher query just isn’t whether or not candidates use AI, however whether or not they use it thoughtfully.
What Separates Good AI-Assisted Functions from Unhealthy Ones
Resume Now’s analysis discovered that 62 p.c of employers reject AI-generated resumes that lack personalization, whereas 78 p.c of hiring managers say customized particulars are what sign real curiosity and match. This tells us one thing necessary: recruiters will not be reacting to AI use itself. They’re reacting to laziness that AI makes simpler.
After reviewing 1000’s of functions, skilled recruiters have a tendency to identify constant patterns that distinguish considerate AI-assisted work from low-effort technology.
Generic functions deal with the job description as a mirror. They mirror the posting’s language again with out including context. The candidate’s abstract reads like a paraphrase of the function necessities somewhat than an account of their precise expertise. Duties are listed however by no means related to outcomes. The quilt letter may apply to any firm in the identical business with a easy find-and-replace of the identify.
Considerate functions, in contrast, present proof of actual engagement with the function. The candidate references particular facets of the crew, the product, or the corporate’s current work. They join their expertise to the job’s challenges in ways in which require precise information of each. Their writing has a perspective — not simply competence, however perspective. These qualities are tough to generate with out real human enter, even with the perfect AI instruments.
The excellence issues as a result of it maps on to the form of worker you’re prone to get. A candidate who takes the time to know your group and articulate a particular worth proposition is demonstrating the identical abilities you have to from them on the job.
Sensible Screening Changes
In case your analysis standards have been designed for a pre-AI utility panorama, they’re overdue for an replace. Listed here are 4 changes that expertise groups can implement with out overhauling their whole course of.
Weight specificity over polish. Practice your reviewers to search for concrete particulars: named tasks, quantified outcomes, references to particular challenges. Polish is now low cost. Specificity nonetheless requires actual information.
Introduce uneven questions. In your utility or early screening, embrace at the least one query that requires the candidate to interact with one thing distinctive to your group — a current initiative, a identified business problem, a hypothetical situation particular to the function. These are tough to reply nicely with a generic immediate.
Transfer analysis downstream. If written functions have gotten much less dependable as a sign, shift extra weight to reside interactions earlier within the course of. A quick telephone display screen or a brief asynchronous video response can reveal communication abilities {that a} polished resume can’t. This doesn’t imply abandoning resume evaluate, however it does imply treating it as one knowledge level amongst a number of somewhat than the first gate.
Consider the work pattern, not the format. For roles that contain writing, evaluation, or technique, contemplate changing the normal cowl letter with a brief work pattern or case response. Give candidates a practical activity and an affordable timeframe. What they produce underneath these situations will inform you excess of any utility doc, no matter the way it was drafted.
The Rising Ability Sign
There’s a broader level right here that’s simple to overlook within the operational particulars. A candidate who makes use of AI instruments successfully — who can take a robust however generic expertise and produce one thing particular, related, and genuine — is demonstrating a talent that the majority organizations desperately want.
The way forward for information work just isn’t about selecting between human effort and AI help. It’s about combining them nicely. Candidates who can do this of their job search are displaying you, in actual time, how they are going to work as soon as they’re in your crew.
This reframes the screening problem. As a substitute of attempting to detect and penalize AI use, the extra productive method is to design analysis processes that floor the qualities AI alone can’t produce: specificity, judgment, real understanding of context, and the power to adapt normal instruments to specific issues.
The organizations that determine this out first is not going to simply enhance their hiring effectivity. They’ll construct groups which are higher outfitted for the best way work truly operates now.
The functions have modified. The query is whether or not our analysis strategies will change with them.
Put up Views: 670
