ai bias in the job search #aitechnology #jobsearch #jobmarket

22 Sept 2025 · 1 min

Watch on YouTube

0:55

In a brief conversational exchange, two speakers discuss an AI‑driven interview assessment platform that ingests video recordings of candidates, analyzes facial expressions, speech, and answers, and then assigns ratings. They note that the system generated backlash because its judgments were not universal—cultural norms, disabilities, speech impediments, or atypical facial expressions caused systematic bias. The company behind the tool ultimately retired it after discovering bias across roughly 19 million analyzed interviews, acknowledging that the AI was predisposed to unfairly penalize certain candidates.

They used AI to do this. It's not universal.
You can't just totally rely on AI because you're gonna have some issues.
There was this company that creates this assessment program that basically ingests these videos and then analyzes people's answers, analyzes their facial expressions and everything, and rates you. They used AI to do this, and they came under a lot of backlash because it's not universal. It can change due to culture or context or disabilities. And so they were finding as a result that the AI was actually predispositioned And judging and rating candidates who maybe had a slight speech impediment. Maybe their expression was a certain way. Maybe they didn't smile as much as the AI wanted it to. So this company actually had to retire this AI tool because, like, 19 million interviews that they had analyzed, but that they found that there was bias as a result.