TestGorilla Report Finds 59% of Firms Made Bad AI Hires

A report from TestGorilla reveals that 59% of organizations in the US and UK made a bad AI hire in the past year. The study, based on a February 2026 survey of 1,928 senior hiring leaders, indicates that 53% of hiring managers now prioritize AI fluency over domain expertise. This shift has created an infrastructure paradox where companies rely on subjective interviews and tool awareness rather than objective skill assessments.

These bad hires are characterized as candidates who can convincingly discuss AI jargon and concepts but lack the practical ability to apply them to their roles. The report notes a significant transatlantic divide, with 33% of US organizations reporting frequent AI-driven errors compared to 13% in the UK.

TestGorilla CEO Wouter Durville attributes these failures to recruitment processes that favor communication skills over execution and lack shared assessment rubrics. To mitigate these risks, Durville recommends implementing structured skills assessments that mirror real work, including tests to identify AI hallucinations. He also suggests conducting skills audits for existing employees to determine if targeted upskilling is a more viable alternative than new hiring.