There is a focus on achieving ai benchmarks, but the accuracy of these measurements is not being sufficiently assessed. The pursuit of ai benchmarks appears to be prioritized over verifying their truthfulness. This disparity raises questions about the validity of these benchmarks.