Report an issue

Practice quizzes for Australian citizenship, driving, and immigration topics

Quizzes

Understanding AI Agents

The agent invents a tool name or arguments that do not exist, potentially causing silent failures or triggering unexpected real-world actions if a matching endpoint exists

The agent calls a real tool but interprets its output incorrectly

The tool returns data that contradicts the agent's training, causing the agent to distrust correct results

The agent calls a tool at an extremely high frequency, overwhelming the external service

What is a 'hallucinated tool call' and why is it a significant risk in agentic systems?