AI Benchmarks Miss 95% of Human Tasks

A study by Mucore shows that AI benchmarks inadequately test 95% of the detailed human tasks across US occupations, leaving major gaps in AI’s real-world capability assessment. This gap risks flawed policy decisions and unpredicted labor market disruptions as current benchmarks focus mostly on a narrow selection of job types and tasks.

This Wire brief sits within Fusion42's coverage of AI & ML.