Analysis
What Is Confused Deputy in AI? Definition & Examples
Publication date unknown · Discovered nhimg.org
Page published
Publication date unknown · First observed: 10 Nov 2026
Coverage timeline
Single-source analysis — one report is available.
Why it matters
Confused-deputy failures can turn an authorized AI agent into a proxy for unauthorized access, making requester-level authorization and tightly scoped tool permissions essential defensive controls.
NHI Mgmt Group’s glossary describes confused-deputy failures in AI as authorization failures where a less-trusted requester steers an agent into misusing its trusted permissions. The reference entry outlines risks from prompt injection, unsafe delegation, and weak tool boundaries, recommending least privilege, context-aware authorization, and separation of requester intent from tool authority.