Social Bias in Popular Question-Answering Benchmarks
This highlights a critical issue for AI fairness, as biased benchmarks can perpetuate unfairness in LLMs, though it is incremental in exposing existing problems rather than proposing new solutions.
The study analyzed 30 QA/RC benchmark papers and 20 datasets, finding that most lack transparency about creators and annotators, with only one addressing social representation, and revealed gender, religion, and geographic biases in the content.
Question-answering (QA) and reading comprehension (RC) benchmarks are essential for assessing the capabilities of large language models (LLMs) in retrieving and reproducing knowledge. However, we demonstrate that popular QA and RC benchmarks are biased and do not cover questions about different demographics or regions in a representative way, potentially due to a lack of diversity of those involved in their creation. We perform a qualitative content analysis of 30 benchmark papers and a quantitative analysis of 20 respective benchmark datasets to learn (1) who is involved in the benchmark creation, (2) how social bias is addressed or prevented, and (3) whether the demographics of the creators and annotators correspond to particular biases in the content. Most analyzed benchmark papers provided insufficient information regarding the stakeholders involved in benchmark creation, particularly the annotators. Notably, just one of the benchmark papers explicitly reported measures taken to address social representation issues. Moreover, the data analysis revealed gender, religion, and geographic biases across a wide range of encyclopedic, commonsense, and scholarly benchmarks. More transparent and bias-aware QA and RC benchmark creation practices are needed to facilitate better scrutiny and incentivize the development of fairer LLMs.