On Benchmarking Human-Like Intelligence in Machines
This addresses the problem of flawed AI evaluation for researchers and developers, though it is incremental as it critiques existing methods without introducing new AI capabilities.
The paper argues that current AI benchmarks are insufficient for assessing human-like intelligence, identifying shortcomings like lack of human-validated labels and simplified tasks, and proposes five recommendations for more rigorous evaluations.
Recent benchmark studies have claimed that AI has approached or even surpassed human-level performances on various cognitive tasks. However, this position paper argues that current AI evaluation paradigms are insufficient for assessing human-like cognitive capabilities. We identify a set of key shortcomings: a lack of human-validated labels, inadequate representation of human response variability and uncertainty, and reliance on simplified and ecologically-invalid tasks. We support our claims by conducting a human evaluation study on ten existing AI benchmarks, suggesting significant biases and flaws in task and label designs. To address these limitations, we propose five concrete recommendations for developing future benchmarks that will enable more rigorous and meaningful evaluations of human-like cognitive capacities in AI with various implications for such AI applications.