AI Agents Push Humans Out of the Loop

A position paper on arXiv argues that increasing AI agent autonomy and extended use degrade human cognitive capacities required for effective oversight.