The Truth About AI Interview Tools: What the Research Really Shows

Everyone's buying AI interview scoring tools — but do they actually work? In this episode, we dig into the independent research on AI interview assessment accuracy, and the findings might surprise you. The short version: most AI interview tools score candidates at validity levels comparable to unstructured phone screens, not the structured behavioral interviews they're marketed as replacing. A massive meta-analysis by Schmidt and Hunter found structured interviews hit validity coefficients of .48 to .51 — independent research on AI tools clusters between .20 and .35. That gap is worth understanding before you sign another contract. We also cover the bias evidence, which is mounting fast. A 2023 University of Melbourne study found statistically significant bias against non-native English speakers across all six AI hiring tools tested — all six. And with only 22 percent of employers actively testing their tools for adverse impact, four out of five organizations are taking on legal risk they may not even know about. The EEOC was clear in 2023: the employer, not the vendor, is liable. Finally, we walk through five questions every HR leader should ask before buying or renewing an AI hiring tool — including what a real independent bias audit looks like, and why candidates are quietly opting out of your process if you're not disclosing AI use.