AI Agents Push Humans Out of the Loop

Summary

Position paper arguing that current agent design does not merely fail to support human oversight but actively degrades it, because the cognitive capacities oversight requires are themselves eroded by extended AI use. Connects the automation and HCI literatures to agent processes and specifies design-level affordances and organizational protocols meant to preserve critical judgement and counteract skill atrophy. The clearest recent statement of the deskilling-plus-oversight problem, and it transfers directly to the question of whether a researcher supervising an agent can still catch its errors.

Contribution

Queued — not yet read.

Method

Queued — not yet read.

Relevance to RISE

Surfaced by the 2026-09-07 literature sweep as a high-priority candidate for the RISE knowledge base (angle: human–AI research collaboration). Verified against: Opened https://arxiv.org/abs/2608.23642, read full abstract, 3 authors (Mitchell, Ghosh, Passi), 2026-08-24, cs.AI/cs.HC. Dedupe: no 'out of the loop' match.

Critique / open questions

Queued — not yet assessed.

Key quotes

AI agents pose significant risks as they are granted increasing autonomy. (abstract)

A commonly proposed solution is human oversight and keeping a ''human in the loop'', but this is not a simple solution: Not only do current approaches to AI agent design impede effective human oversight, but the cognitive capacities required for it are also themselves degraded by extended use of AI systems. (abstract)