Rumination on automation, 1998

摘要

In recent years there has been great interest in the evidence that with experience human decision-making becomes "automatic", "recognition-primed", "naturalistic", "situated", and "ecological". Advocates of these approaches have rejected classical decision theory as a description of decision-making in real tasks. This rejection poses a challenge to better understand human automaticity by extending normative decision modeling. Two approaches, both based on state-space reduction, are discussed. There is also the challenge of relating these models to what is being learned in neurobiology, and again a model is presented which potentially explains how behavior can become automatic. Implications for computer-based automation of decision aiding and control are also proposed, including a revised four-stage format for scaling levels of automation.