AI assistant functions influence trust differently based on user knowledge
How Far Do Capability Cues Travel? Anthropomorphism and Differentiated Trust in a Platform-Embedded AI Assistant
Human-Computer Interaction
Summary
People don't always trust AI assistants just because they see the AI can do more things. The authors showed that when an assistant displayed three functions instead of one, users perceived it as more multifunctional, but this didn't always increase trust in intelligence. Giving explanations for the functions didn't make users think the AI was smarter. People who knew more about AI responded more strongly to seeing more functions. How users view the AI’s intelligence, anthropomorphism (seeing it as human-like), and trust factors like benevolence, integrity, and ability are related but different.
What this means in practice
- •For product designers: Design AI assistant interfaces that show multiple functions to enhance perceived multifunctionality for users with higher AI literacy.
- •For ux researchers: Evaluate trust components—ability, integrity, and benevolence—separately when studying user trust in AI assistants embedded in platforms.
Authors
Chenchen Mao, Hanjing Shi, Haiyan Jia, Dominic DiFranzo
Abstract
Visible AI capabilities need not translate into broader judgments of trustworthiness. In a randomized 2 x 2 experiment with 270 U.S.-based Reddit users, an embedded assistant displayed one or three functions, with or without a brief rationale. Displaying three functions increased perceived multifunctionality; no other randomized main effect survived correction across the six outcomes. Rationale availability did not reliably increase perceived intelligence. Exploratory analysis indicated stronger uptake of the functional display at higher objective AI literacy. Among concurrently measured judgments, perceived multifunctionality was associated with perceived intelligence, which was associated with anthropomorphism and all three trust dimensions. After accounting for perceived intelligence, anthropomorphism was positively associated with benevolence, but not reliably with integrity or ability. These findings separate interface effects from relationships among users' perceptions and show why ability, integrity, and benevolence should be evaluated separately.