A risk analyst is evaluating an AI system in which multiple LLM agents collaborate: a planning agent decomposes tasks, execution agents carry them out, and a critic agent reviews outputs. Which failure mode is MOST unique to this multi-agent collaborative architecture?
Select an answer to reveal the explanation.
Short Explanation
You want the failure mode that comes from agents depending on each other. If the planning agent breaks the task down wrong, the execution agents can all follow that bad plan without checking it independently, so one planning mistake spreads through the whole workflow.
Full Explanation
The correct failure mode is the planning agent producing an incorrect decomposition that all execution agents follow without independent verification. This is unique to multi-agent collaboration because agents create correlated failure risk: one planning error is inherited by downstream execution agents, amplifying it across the workflow. Running out of context window capacity is a technical limitation. Training the critic on different data is a secondary misalignment concern. Requiring more API calls and increasing latency is a performance characteristic, not a failure mode.