The evidence behind the guides
Sources Find a study, check what it supports, and follow it back to a concern.
TopicAll topics Data Integrity & Provenance Model Performance & Reliability Security & Adversarial Risk Transparency, Explainability & Accountability Agentic & Multi-Agent Risks Infrastructure, Compute & Deployment Environmental & Physical Footprint Discrimination, Fairness & Toxicity Privacy & Surveillance Misinformation, Manipulation & Democracy Malicious Use & Weaponization Human–Computer Interaction Harms Health, Safety & Care Labor, Work & the Economy Education & Knowledge Inequality, Access & Power Culture, Identity & Values Governance & Regulation Rights, Law & Justice Corporate & Enterprise Concerns Power Concentration & Geopolitics AI System Safety & Alignment Catastrophic & Existential Risk CollectionAll sources Cited in concern guides Topic reading Research tools & guidance
Sources for Confused-deputy attacks Show all concerns 3 sources
Guide citations are still under editorial review. Greshake and colleagues · 2023 · Research study
Demonstrates indirect instructions through retrieved content in specified applications and prototypes.
Evidence notes & 6 linked concerns + Internet Engineering Task Force · 2025 · Security standard
Defines current OAuth security practices, including limited privileges and intended-resource checks; implementations still need verification.
Evidence notes & 5 linked concerns + Model Context Protocol maintainers · July 2026 documentation · Official technical guidance
Explains confused-deputy and token-passthrough risks in MCP implementations; not proof that every connected agent has these defects.
Evidence notes & 2 linked concerns +