Where do you stand?

24 questions, about 18 minutes — 3 from each module. Answer honestly and skip what you don't know: we'll build a week-by-week study plan around your gaps.

You can take the diagnostic as a guest. Create a free account to save your plan and track progress against it.

Your plan settings

With an exam date we fit the plan to the weeks you have and tell you if you need more hours.

Module 1 · AI & GenAI Foundations for Security Professionals

Question 1 Question 1 of 24
What does LoRA do?
Question 2 Question 2 of 24
An agent stores a summary of each conversation in long-term memory. An attacker plants instructions in one conversation that alter the agent's behaviour in later, unrelated sessions. This is:
Question 3 Question 3 of 24
Confidential computing using trusted execution environments primarily protects data:

Module 2 · AI Governance, Regulation & Responsible AI

Question 4 Question 4 of 24
A deployer substantially modifies a high-risk AI system and markets it under its own brand. What is the MOST significant consequence under the EU AI Act?
Question 5 Question 5 of 24
Network logs show 1,400 staff using public chatbots with corporate data. What is the MOST effective FIRST response?
Question 6 Question 6 of 24
A declined loan applicant asks why. Which explanation type is MOST useful to them?

Module 3 · AI Risk Management & the AI Supply Chain

Question 7 Question 7 of 24
Which layer is the BEST place for least-privilege scopes and approval gates?
Question 8 Question 8 of 24
A public AI assistant receives about 200 injection attempts a year, and red-team results suggest 2% would succeed. What is the expected loss event frequency?
Question 9 Question 9 of 24
A team wants to use a LoRA adapter, saved in a pickle-based format, from an unknown publisher. What is the MOST significant risk?

Module 4 · Data Governance, Privacy & Protection for AI

Question 10 Question 10 of 24
Why is cryptographic erasure often preferred for sanitising AI data in cloud storage?
Question 11 Question 11 of 24
A consent form says 'we may use your data to improve our services and AI'. What is the MAIN weakness?
Question 12 Question 12 of 24
Which technique helps find training records with outsized influence on a model's behaviour?

Module 5 · Model Security & the Secure AI Lifecycle

Question 13 Question 13 of 24
Which change typically warrants the deepest review?
Question 14 Question 14 of 24
An attacker steals a data scientist's notebook token. Which design MOST limits the damage?
Question 15 Question 15 of 24
In model risk management, what is independent validation?

Module 6 · Secure AI Architecture, Guardrails & Agentic IAM

Question 16 Question 16 of 24
Which restriction is MOST important for a sandbox where an agent runs code it wrote?
Question 17 Question 17 of 24
In the A2A protocol, what does an Agent Card describe?
Question 18 Question 18 of 24
Which tool design is MOST secure for an assistant that answers order questions?

Module 7 · AI Threat Modelling, Red Teaming & Assurance

Question 19 Question 19 of 24
What is MAESTRO designed for?
Question 20 Question 20 of 24
How should a mitigation be validated?
Question 21 Question 21 of 24
Why should model, prompt and guardrail versions be pinned during a red-team engagement?

Module 8 · AI Security Operations, Incident Response & Recovery

Question 22 Question 22 of 24
What is the difference between a playbook and a runbook?
Question 23 Question 23 of 24
Before rolling back a suspected tampered model, what should responders do FIRST?
Question 24 Question 24 of 24
Which component is the BEST collection point for structured tool-call telemetry?