73% of students and graduates have used AI in their job applications only 38% think it's fair for employers to screen CVs with AI (spot the problem)

that's from an Institute of Student Employers survey published in June.

AI in hiring is now an arms race where both sides use the same tools and neither trusts the other's.

so here's the view from both desks: the graduate applying, and the recruiter screening.

How AI in hiring looks for graduates

from the applicant's side:

  1. AI writes the application. 46% of students in the ISE survey used AI to write a CV or cover letter, and 95% believe employers use AI to shortlist
  2. AI runs the interview. a Greenhouse survey of US job seekers in May found 63% had been interviewed by an AI, and 70% weren't clearly told upfront that AI would evaluate them. of those who finished an AI interview, 51% never heard back
  3. the market is tight anyway. the New York Fed puts unemployment for US graduates aged 22 to 27 at 5.7% as of June (CNBC)

(Greenhouse sells AI interviewing tools, so read its numbers with that in mind)

How AI recruitment looks for recruiters

now flip the desk:

→ 67% of US HR leaders told Robert Half that AI-generated applications have slowed their hiring, and 65% said the surge makes it harder to verify skills → 38% now interview each candidate more times, and 32% are rewriting job descriptions to discourage generic AI answers → LinkedIn research found 66% of recruiters planned to expand their use of AI for pre-screening interviews at the start of 2026 (CNBC)

so the loop goes: candidates use AI to apply more → recruiters get buried → recruiters use AI to screen → candidates use AI to beat the screen.

honestly nobody wins that loop.

The legal side is catching up

in June US District Judge Rita Lin refused to dismiss claims that Workday's AI screening tools discriminated against applicants on disability, race, age and gender (HRD).

one surviving claim is that the software flags proxies for disability, like gaps in employment history.

Workday denies wrongdoing and says its technology only assesses job qualifications.

either way, the tools making the first cut are now being tested in court.

What uni won't teach you about AI and hiring

careers services mostly teach the CV as a document.

now it's an input to a filter, read first by software and then (maybe) by a very tired human.

→ bad: a polished AI-written cover letter that reads exactly like everyone else's → good: one specific line only you could write, about something you actually built or did

and if you're heading into HR, nobody teaches you how to question a screening tool. that's quickly becoming part of the job.

The critical-thinking move: 5 questions for both sides of AI hiring

for graduates:

  1. what is the filter looking for? read the job description like a spec and match your real skills to it (stuffing in skills you don't have just fails at interview)
  2. what proves this is me? a link to a project, a number from a real result. something an AI couldn't make up for you
  3. am I applying better, or just more? the ISE found AI use was highest among people sending 101 to 200 applications. volume isn't a strategy

for recruiters:

  1. what is the tool actually scoring? if you can't explain why a candidate was rejected, you can't defend it (ask Workday)
  2. was the candidate told? 70% of Greenhouse's respondents weren't clearly told AI was evaluating them. disclosure costs nothing and builds trust

What to do this month

  1. take one application you've sent and delete every sentence that could appear in anyone else's. replace them with one specific example
  2. if an employer uses an AI interview, ask beforehand what it measures and whether a human reviews the result
  3. HR students: find a target employer's AI recruitment policy and test it against the five questions above
  4. record one mock interview on camera with no AI help at all, then watch it back. that's the version of you both sides are trying to find

so TLDR: AI made applying easy and screening hard, so everyone is now judged on the one thing AI can't fake: real evidence.

are you using AI to apply better, or just to apply more?

Sources

Get the next issue by email

theVANTAGEd is written by students to bring critical thinking to AI across real industries. One issue every two weeks.