Detecting moment-specific cognitive demand for adaptive AI interaction

Determine how an AI system for data-literacy assistance can detect, at any given interaction moment, whether a user’s cognitive demand is receptive (seeking information or confirmation) or deliberative (requiring scaffolded reasoning), and adapt the system’s interaction mode accordingly (transmissive versus interrogative) to maintain cognitive alignment.

Background

The paper argues that AI’s default one-off, transmissive responses can induce cognitive passivity, especially for novices, and proposes a cognitive alignment framework matching users’ cognitive demand (receptive vs. deliberative) with AI interaction modes (transmissive vs. interrogative).

A key open challenge on the AI side is dynamically inferring what the current moment demands so the system can adjust its behavior in real time, avoiding misalignment that leads to cognitive passivity or friction.

References

While this paper aims to discuss the high-level notion for advancing AI-assisted literacy, we leave questions for the community to answer in the future research: (1) From the AI side, how can a system detect what a given moment demands, and adapt accordingly?

— Disrupting Cognitive Passivity: Rethinking AI-Assisted Data Literacy through Cognitive Alignment  (2604.02783 - Ahn et al., 3 Apr 2026) in Section 1, Introduction