Defending Against Agent Abuse

Final Quiz

Connecting to LMS... Progress: in progress

Version 1.0 | Date: 2026-06-21

This course provides educational information regarding defensive security controls for AI agents. It is not offensive security training, exploit development instruction, or legal advice.

Final Quiz

1. Why can AI agents present significant organizational risk? A. They may access information, tools, and perform actions. B. They are always isolated from business systems. C. They cannot affect real environments. D. They eliminate security concerns.

2. Which component may be part of an agent attack surface? A. Tool integrations and external content. B. Office decorations. C. Marketing slogans. D. Parking assignments.

3. What does least privilege mean for AI agents? A. Agents receive only the permissions needed for approved tasks. B. Agents receive unrestricted administrator access. C. Agents can access every available system. D. Agents do not require permission controls.

4. Why should agents treat external content carefully? A. External content may contain misleading or malicious instructions. B. External content is always trustworthy. C. External content eliminates security concerns. D. External content replaces governance.

5. Which activity may require human approval? A. High-impact operational changes. B. Reading public documentation. C. Displaying help information. D. Opening a dashboard.

6. Why should organizations log agent activity? A. To support accountability, investigations, and improvement. B. Logging is unnecessary. C. Logging eliminates all risk. D. Logging replaces security controls.

7. Why should agents undergo continuous validation? A. Risks and behaviors may change over time. B. Initial testing eliminates future risk. C. Validation is unnecessary after deployment. D. Validation replaces monitoring.

8. Why is governance important for AI agents? A. Governance establishes policies, accountability, and oversight. B. Governance eliminates analytical errors. C. Governance removes the need for monitoring. D. Governance replaces security controls.

9. Which statement best summarizes defenses against agent abuse? A. Effective defenses combine technical controls, governance, and human oversight. B. Unlimited agent autonomy is always safe. C. Monitoring is unnecessary. D. Governance is optional.

10. What security principle helps reduce the impact of agent misuse? A. Least privilege. B. Unlimited access. C. No monitoring. D. No approval processes.

11. Why are approval gates useful? A. They require human review before high-impact actions. B. They eliminate governance requirements. C. They allow unrestricted automation. D. They remove accountability.

12. Which activity supports agent accountability? A. Logging important actions and tool usage. B. Disabling monitoring. C. Removing audit trails. D. Ignoring agent behavior.

13. Why should organizations continuously test agents? A. Agent behavior and risks may evolve over time. B. One test is sufficient forever. C. Testing is only needed after incidents. D. Monitoring replaces testing.

14. Which statement is most accurate? A. Defending against agent abuse requires layered controls and governance. B. AI agents are safe by default. C. Agent monitoring is unnecessary. D. Security controls limit all business value.