Attack and defend an AI assistant. You work as the AI red team for a fictional university, break its student-help assistant in your own private lab, find your attacks in the logs, add safeguards, and prove the fixes hold. You finish with an AI security assessment report you can keep as proof of your work.
Includes the full course, your private Ask Odie lab, lifetime course access and a certificate when you pass the final knowledge check and the capstone. The course is self-paced: start any time and work through it on your schedule.
Who it’s for
Security, IT and software people who want to test and secure AI apps and agents: SOC analysts, penetration testers, cloud security engineers and security leaders. No AI background needed.
What you’ll do
- Break. Attack Ask Odie with prompt injection, data leaks, poisoned documents, unsafe output and misused tools.
- Catch. Find your own attacks in the logs and write detections.
- Fix. Turn on the right safeguards and understand why each one works.
- Prove. Re-run the attack and show it now fails, with a regression suite.
- Report. Write an assessment report for the university’s CIO in an AI-graded capstone.
Aligned with DoD 8140 work role 733 AI Risk & Ethics Specialist.