Does Higher Algorithm Adaptability Increase Usage Intention Among Professional Users? A Dual-Path Mechanism Moderated by Task Type

As adaptive algorithms permeate professional settings, does higher algorithm adaptability enhance professional users’ usage intention? Drawing on Self-Determination Theory, three complementary studies were conducted: secondary analysis of Doubao AI-assistant reviews (N = 409), a between-subjects experiment with 190 participants simulating language-editing tasks, and a replication with 193 university faculty in authentic teaching contexts. Results showed that adaptability enhanced usage intention in subjective tasks but attenuated it in objective tasks. Mechanistically, higher adaptability promotes usage intention by strengthening users’ confidence in their ability to perform the task (task self-efficacy); however, in objective tasks it also erodes perceived transparency—a precondition for professionals to verify and take accountability for algorithmic outputs—thereby dampening usage intention. These opposing pathways jointly weaken the effect in objective tasks. This study extends algorithm-attitude theory by treating adaptability as an inherent design attribute, clarifying its dual pathways and boundary conditions, with practical implications for professional AI design.