Understanding Barriers to <scp>AI</scp> Use Among Public Sector Knowledge Workers

ABSTRACT Artificial intelligence (AI) promises to streamline workplace information practices, yet little is known about its adoption in public‐sector contexts. We surveyed 103 Israeli public‐sector knowledge workers regarding: (a) their willingness to use AI for 16 work‐related information activities and (b) their actual use of such tools. Principal‐component analysis grouped the activities into three categories: strategic &amp; analytical tasks, management tasks, and personal applications. Across all categories, paired‐t tests showed a large willingness–use gap (e.g., strategic tasks: M = 3.67 vs. 1.88, p &lt; .001). Multivariate regression revealed that AI availability strongly predicted use ( η 2 ≈ .14–.22) but weakly predicted willingness ( η 2 ≈ .04–.05). Results suggest that providing access is necessary but not sufficient; motivational and organizational factors moderate adoption. We outline design and policy steps, including trust‐building, role‐aligned training, and low‐risk trial contexts, to bridge the gap and advance human‐centered AI in the public sector.