Research Summary: AI-Augmented Inquiry and Regulation in Hybrid Systems: A Control Allocation Architecture for Preserving Epistemic Agency in Hybrid Human-AI Cognition

The integration of Generative Artificial Intelligence (GenAI) into epistemic processes, such as hypothesis generation and decision-making, while enhancing performance, introduces a significant metacognitive dilemma. This dilemma manifests as a decline in internal monitoring and cognitive engagement as external generative capacity increases, representing a redistribution of cognitive control in human-AI systems. The AIRIS (AI-Augmented Inquiry and Regulation in Hybrid Systems) framework is proposed to analyze this issue and identify regulatory intervention points.

Why it matters

This development highlights critical challenges in maintaining human cognitive capabilities and agency as advanced AI systems become more pervasive in knowledge work and decision-making. Addressing the metacognitive dilemma is crucial for ensuring effective human oversight and maintaining high-quality outcomes in AI-augmented environments, preventing unintended skill degradation or over-reliance.

Key insights

• Generative AI systems are increasingly being integrated into epistemic processes.
• GenAI reliably enhances performance in areas like hypothesis generation, explanation construction, and decision-making.
• A metacognitive dilemma arises where increased external generative capacity correlates with a decline in internal monitoring, calibration, and cognitive engagement.
• This dilemma signifies a redistribution of cognitive control within distributed human-AI systems.
• The phenomenon is distinct from traditional automation bias or mere reliance on algorithms.
• The AIRIS framework is introduced to analyze this dilemma and specify regulatory intervention strategies.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.21618