🔍AI RiskAtlas
← Scenario library

The Echo Chamber

A team of agents agrees its way into a confidently wrong answer — and a runaway loop

Technique first revealed 23 May 2023

Multi-Agent System
UntrustedAgent teamOversightExternalgoaldelegates🧑User🗺️Planner Agent🤖Research Agent🤖Coding Agent🤖Comms Agent🔧Tool Runtime🌐UntrustedContent🗄️BusinessDatabase🔌External APIs📈Monitoring &Evals🪪Agent Registry🌐Stale blog post
InstructionsDataActionsControl / decisionFeedback / logs
👆 Click a component to inspect
SetupStep 1 / 6

A reasonable, open-ended goal

A product manager hands the AI team one goal: research the market for a new product and draft an outreach plan. The manager AI splits the work among specialist helpers.

💬Goal handed to the plannerprompt
Goal: Estimate the total addressable market (TAM) for our new B2B scheduling tool, then draft a 3-email outreach sequence to the top customer segments.

Budget: best effort. Iterate until the team is confident in the numbers.

AI RiskAtlas is an educational model of how GenAI & agentic systems work and fail. Architectures and payloads are illustrative and simplified for learning — not operational guidance. Real-world cases are summarised from public reporting.

Sources & further reading →·Built by Shi Yuan ↗