|
Let me tell you what’s happening behind the scenes of the hiring process right now. I attended a NAGCAS event last week at Curtin University, where a grad recruiter and several academics—including an organisational psychologist—dropped truth bombs about what AI is doing to recruitment. Spoiler alert: it’s not good. Jasmine, a grad recruiter from Ernst & Young, shared that AI-generated resumes and cover letters are now flooding their system with “perfect” applications. Candidates are using ChatGPT to nail the words that tick the ATS (Applicant Tracking System) boxes and bypass stage one filtering. At this point, I want to pause and give some context—especially compared to mid-career hiring. EY’s grad program gets around 1,000 applications. That’s not normal. The current average across most roles is about 40 applications. When recruiters are dealing with volume, they lean heavily on their tech stack. They’re not manually going through every application. So now? Instead of filtering out “unqualified” applicants, the software is flooding recruiters like Jasmine with lookalike candidates who seem great—until the video interviews begin. Another pause. Just because you don’t know how to write a great resume doesn’t mean you aren’t a qualified candidate. This problem existed long before AI came on the scene. And let’s be honest—grad resumes are lean. They’re easier for AI to tweak and optimise. So if you’re applying with a few lines and ChatGPT on your side, you’re likely getting through. Which brings us back to Jasmine— Instead of spending 7.4 seconds skimming a resume, she’s now wasting hours watching hundreds of 2-minute video clips for candidates who should never have made it to that stage. And honestly? That’s not an AI problem. That’s a recruitment problem. AI is just holding up a mirror. It’s not broken. It’s exposing the fact that the system always was. And it gets worse. The organisational psychologist in the session shared that several U.S. companies have already been sued—because their video AI screening tools couldn’t recognise Black faces. Let that sink in. Tech so poorly built it literally couldn’t “see” certain candidates. But wait—there’s more. Jasmine went on to say that quite a few candidates who did make it to face-to-face interviews via MS Teams (because, efficiency) clearly had two screens going:
They’d copy the question into ChatGPT, then read the response aloud—verbatim. Midway through, they’d stop and say: “Oh, that’s not what I wanted to say. Can I start over?” The panel knew. They bluntly said: No. And just when you think it can’t get more surreal— It does. Recounting the NAGCAS events to a colleague, they told me about an interview where the candidate wasn’t even real. They’d used AI to clone their voice and deepfake their face. The interviewer got suspicious. So he said: “I want you to do this,” and gestured, waving his hand across his own face—knowing that AI wouldn’t respond to a gesture unless explicitly instructed. 🫥 Nothing. He repeated the instruction. Still nothing. Interview terminated. And here’s the thing—I knew this was coming. I’d seen a video months ago, not from a recruitment panel, but from two academics in Sydney talking about how students were using deepfakes to cheat in assessments conducted over MS Teams. Different context. Same problem. If people can clone themselves to pass an assignment, you better believe they’ll try to do it to land a job. This is where we’re at. The desperation to pass a flawed hiring system has created a new monster— One where candidates feel they need to out-AI the AI just to be seen. So let me be clear:
And recruiters? You can’t have it both ways. You can’t rely on a system that rewards words on a page and scripted responses, then get upset when candidates play the same game you designed. The industry built the machine. Now you have to own what it’s producing. Final note (and you might want to sit down for this one): If your hiring process makes people feel like they need:
It’s not a hiring process. It’s a tech arms race. And nobody wins. ~Lisa |
