AI agents (agentic AI)
AI systems that plan and take actions using tools, not just produce text.
Agents can call APIs, run code, send emails or change records to reach a goal with limited human input. This raises the stakes of errors and prompt injection, because outputs become actions.
On the exam
Key controls: least-privilege tool access, approval gates for sensitive actions, activity logging and the ability to stop the agent.
Related terms
- Human in the loopA person reviews or can override an AI system’s decisions.
- Prompt injectionText that tricks a language model into ignoring its instructions.
- Least privilegeGiving each user only the access their job needs.
AAIA Test yourself on this with timed AAIA practice questions in the AAIA Exam Simulation Hub, or read the free AAIA exam guide.
Test yourself with today’s question of the day.
