# Position: Evaluating Bias in AI Agents Is (Still) a Social Science Measurement Challenge

Status: forthcoming
Visibility: public-pdf
Status label: Forthcoming, ACL 2026 Workshop on Evaluating AI in Practice
Authors: with Soubhik Barari and Diwas Bhattarai
Canonical URL: https://openreview.net/forum?id=gYaR4LxoBa

## Summary

This position paper argues that evaluating bias in AI agents is still a social-science measurement challenge. It distinguishes dimensions of bias, separates observed bias from its causes, and highlights delegation as a source of bias in agentic systems.

## Public Files

- Metadata: https://jyl19.github.io/papers/position-evaluating-bias-in-ai-agents-is-still-a-social-science-measurement-challenge/metadata.json
- PDF: https://jyl19.github.io/papers/position-evaluating-bias-in-ai-agents-is-still-a-social-science-measurement-challenge/paper.pdf