Boosting metacognition in entangled human-AI interaction to navigate cognitive-behavioral drift

TL;DR

The paper addresses how entangled interactions with adaptive AI, especially LLM-based chatbots, can gradually shape human cognition and behavior through drift. It proposes a framework centered on three core phenomena—entanglement, cognitive/behavioral drift, and metacognition—and identifies four metacognitive intervention points to scaffold user judgment and action: interaction initiation/role gating, confidence cue calibration, drift detection, and verification-based gating. It discusses how these dynamics unfold across micro, meso, and macro levels, including potential impacts on democratic discourse and public health if left unmanaged. A forward-looking research agenda is offered, detailing drivers, measurement, intervention testing across multiple levels, scaling approaches, and policy considerations under non-cooperation to preserve epistemic agency in AI-mediated environments.

Abstract

People navigate complex environments using cues, heuristics, and other strategies, which are often adaptive in stable settings. However, as AI increasingly permeates society's information environments, those become more adaptive and evolving: LLM-based chatbots participate in extended interaction, maintain conversational histories, mirror social cues, and can hypercustomize responses, thereby shaping not only what information is accessed but how questions are framed, how evidence is interpreted, and when action feels warranted. Here we propose a framework for sustained human-AI interaction that rests on invariant features of human cognition and human--AI interaction and centers on three interlinked phenomena: entanglement between users and AI systems, the emergence of cognitive and behavioral drift over repeated interactions, and the role of metacognition in the awareness and regulation of these dynamics. As conversational agents provide cues (e.g., fluency, coherence, responsiveness) that people treat as informative, subjective confidence and action readiness may increase without corresponding gains in epistemic reliability, making drift difficult to detect and correct. We describe these dynamics across micro-, meso-, and macro-levels. The framework identifies four metacognitive intervention points and psychologically informed interventions that provide metacognitive scaffolding (boosting and self-nudging). Finally, we outline a long-horizon research agenda for scientific foresight.

Figures (1)

• Figure 1: Entanglement in human--AI interactions. (A): Entangled interaction between user and AI and cognitive drift leading to behavioral drift. (B): Three levels of analysis: At the micro level, people interact individually with their respective GenAI systems; at the meso level, small groups of people develop individual entanglements with their respective AI, but they will also influence each other and even people who are not interacting with AI; at the macro level the entanglement influences whole populations. (C): Expanding on the meso level: how behavioral drift influences other users. (D): Illustration of the four metacognitive intervention points on the entangled interaction.