US
AU
UK
NZ
Other countries
The Hidden Dangers of AI “Cheating Detection” in Hiring
AI-powered hiring tools are being marketed with bold claims, with some boasting “90% accuracy at detecting cheating in video interviews.” At first glance, that sounds impressive. But look closer: a 10% error rate means 1 in 10 innocent candidates could be falsely branded as cheats. At scale, that’s not just a flaw—it’s a disaster.
Why False Positives Are So Harmful
-
Career damage: Wrongly accused candidates may lose opportunities they’ve worked hard for.
-
Employer brand erosion: Companies risk being seen as unfair or careless in their hiring.
-
Trust breakdown: Both candidates and employees lose faith in the process.
False positives don’t just filter out “bad actors”, they exclude talented people in a world where talented people are hard to find.
The Snake Oil Problem
Vendors claim they can detect “cheating” by analyzing voice tone, facial movements and expressions, or whether someone is reading from a script. But there’s no credible scientific basis for this. These systems are destined to get it wrong at scale, harming people who’ve done nothing wrong and exposing your business to reputational risk.
We’ve seen this before: in 2021, HireVue abandoned its facial analysis feature after sustained criticism over bias risks tied to ethnicity, neurodiversity, disability, and socio-economic background. Yet, the HR tech market is now flooded with shiny “AI-powered” tools, many little more than ChatGPT wrappers hooked up to an external, ever-changing model they have no control over, making ever-bolder claims about automated candidate evaluation.
It’s the Wild West out there, but regulation Is catching up:
-
California: On October 1st 2025, new Fair Employment & Housing regulations outlawed (among other things) AI-based facial expression assessments.
-
European Union: EU Artificial Intelligence Act also lists similar prohibited AI applications in employment.
But until statutory regulation takes hold across more states and countries, these tools risk being slipped into hiring pipelines unchecked.
The Only Test That Matters
At the end of the day, hiring technology should be judged on one standard: does it give every candidate a fair shot at demonstrating their talent and potential?
If a tool can’t prove its accuracy, fairness, and freedom from bias, it has no place in your hiring process. Anything less isn’t innovation; it’s exploitation.
