---
id: "quote-oversight-capacity"
type: "quote"
source_timestamps: ["§ 1. Explicitly redefine workflows and then name new (human) role expectations."]
tags: ["management", "cognitive-load"]
related: ["concept-oversight-capacity"]
speaker: "Matthew Kropp et al."
speakers: ["Matthew Kropp", "Julie Bedard", "Emma Wiles", "Megan Hsu", "Lisa Krayer"]
quote: "Oversight capacity does not expand automatically just because output does."
sources: ["agentic"]
sourceVaultSlug: "hbr-seg-agentic"
originDay: 6
articleStem: "hbr-ext-16-dont-treat-agents-like-employees"
sourceUrl: "https://hbr.org/2026/05/research-why-you-shouldnt-treat-ai-agents-like-employees"
sourceTitle: "Research: Why You Shouldn’t Treat AI Agents Like Employees"
---
# Oversight Capacity Does Not Expand

> "Oversight capacity does not expand automatically just because output does."
> — [[entity-matthew-kropp]] et al.

The authors' single most important design principle. It defines [[concept-oversight-capacity]] as a hard cognitive constraint and warns against the fallacy that 10× AI output permits 10× human oversight. Ignore it and you get [[concept-ai-brain-fry]] and the error spikes in [[claim-brain-fry-errors]]. It anchors Step 1 of the [[framework-responsible-human-ai-collaboration]] and the action [[action-redefine-spans-of-control]].