Dual-process theory — the account of cognition as operating through two interacting systems: a fast, automatic, intuitive System 1 and a slower, effortful, analytical System 2. In education, dual-process theory frames when learners rely on quick heuristics versus deliberate reasoning, and why they sometimes take cognitive shortcuts.
Dual-process theory has strong explanatory value for AI in education. Much of the risk of Over Reliance and Cognitive Offloading can be read as learners defaulting to System 1 (accept the fluent AI answer) instead of engaging System 2 (verify, analyze, reason). Conversely, Desirable Difficulties and desirable difficulties require System 2 engagement.
Relevance to AI in education
Explaining over-reliance: an AI's fluent, plausible output invites System 1 acceptance; unverified trust is a heuristic shortcut. This links to Cognitive Offloading and the "speedup illusion" in human-AI interaction.Designing for System 2: scaffolds that force verification, articulation, and reasoning (draft-first, hint-not-answer, explanation requirements) push learners toward deliberate processing and protect learning.Measuring cognition: dual-process distinctions inform how we interpret learner behavior and performance, and caution against mistaking fluent output (System 1) for understanding (System 2) — a theme in Educational Measurement and Psychometrically Aware AI.Connections
Dual-process theory connects to Cognitive Load Theory, Metacognition, and Cognitive Offloading. It underpins the rationale for Reducing AI Misuse interventions and for assessment designs that distinguish surface fluency from deep understanding.
Connected Concepts
Cognitive Load TheoryMetacognitionCognitive OffloadingOver RelianceEducational MeasurementReducing AI MisusePsychometrically Aware AIConnected Articles
Cognitive Offloading Speedup Illusion — Cognitive Offloading and the Speedup IllusionAI Making Us Stupid — Is AI Making Us Stupid?Efficiency Gain Illusion AI Overreliance — The Efficiency-Gain IllusionLLM Reasoning Traces Metacognition — LLM Reasoning Traces and MetacognitionCognitive Commons AI Expertise Regeneration — The Tragedy of the Cognitive Commons