Task-Aligned vs. Human-Aligned: Why AI’s Next Benchmark Should Be Us

Stanford's 2026 AI Index Report shows rapid progress in AI models on task-based benchmarks, but these do not measure the impact on users. The article argues for shifting focus from task-alignment to human-alignment, evaluating how AI affects users' capabilities after interaction, especially for subjective or complex queries without single correct answers.