How AI Literacy Shapes GenAI Use

Why Read This from an HCI Perspective

This article reframes genAI use not as a simple matter of ‘whether people use it,’ but as a question of what they ask, how they ask it, and how they verify the results. For HCI/UX practitioners, it highlights that prompt proficiency and the ability to judge outputs are distinct design challenges. For researchers, it provides grounds for not treating AI literacy as a single, one-dimensional metric. This is especially relevant when designing for users with low AI literacy and for trust- and verification-focused UX.

CIT's Commentary

From a CIT perspective, the key point of this article is that ‘using AI well’ does not automatically mean ‘judging AI well.’ Many organizations focus on prompt training alone, but the real risk lies in how well users can perceive the accuracy, timeliness, and evidential grounding of the output. Therefore, CIT argues that AI experience design should separate input support from verification support. For example, beyond interfaces that help users ask better questions, we need mechanisms that reduce the cost of judgment—such as tracking core claims, comparing sources, and warning users about potentially outdated information. Also, ‘distrust’ and ‘indifference’ may not be mere resistance; they can be sophisticated usage strategies. As a result, user segmentation should be redrawn around trust and verification behaviors rather than around skill level alone.

Questions to Consider While Reading

• Q.In our product, at what moments do users become ‘naive power users’—people who write prompts well but do not verify the outputs?
• Q.In contexts where output accuracy matters, what UI patterns can naturally lead users to check sources and re-verify?
• Q.When measuring AI literacy, how should prompt proficiency and output judgment ability be separated and evaluated?