Studying People to Study AI: Expert Perspectives on the Epistemic Fit and Barriers of Human Research in AI Safety & Ethics

2026-08-06

AI Summary

This study addresses the overreliance on technical assessments in current AI safety and ethics research, which has led to insufficient understanding of human-AI interaction risks due to the neglect of empirical human-centered studies. Drawing on 93 expert surveys and 17 in-depth interviews spanning four communities—technical, socio-technical, governance, and normative—the research uncovers methodological tensions among scholars with different backgrounds, particularly highlighting the low engagement of technically oriented researchers with human-centered approaches. While the value of human research is widely acknowledged, its integration remains constrained by concerns about validity, scarce resources, methodological preferences, and inadequate infrastructural support. To counter superficial adoption—termed “human-washing”—the paper proposes actionable pathways for effectively incorporating human-centered research into AI safety practices, offering methodological guidance for interdisciplinary convergence.

Abstract

Safety risks of AI are becoming increasingly evident in human interactions with AI technologies. The prominent approaches to evaluating these risks favor technical methods, such as model benchmarks and LLM simulations, often sidelining empirical research with human subjects. To examine this apparent gap in the acceptance of human research, we conduct an expert survey (n=93) and expert interviews (n=17) with AI Safety & Ethics (AISE) researchers from Technical, Sociotechnical, Governance, and Normative backgrounds. Our findings suggest that although there is a consensus that human research is valuable for generating evidence for AISE, its adoption and acceptance are constrained by perceived validity issues, tangible resource barriers, epistemic and personal preferences in methods, and infrastructural constraints from the broader research community. In particular, Technical researchers tend to value human research less and collaborate across disciplines less, suggesting an epistemic tension towards human methods. We propose recommendations for establishing the epistemic fit of human research within AISE and bridging the prohibitive limitations that researchers face, while avoiding performative 'human-washing'.

Problem

• AI Safety
• Human Research
• Epistemic Fit
• Research Barriers
• AI Ethics

Innovation

• epistemic fit
• human research
• AI safety & ethics
• interdisciplinary collaboration
• methodological barriers