University of Guelph · Awarded Jul 3, 2026 · 3-day micro-credential, approx. 15 learner hours
Focus
- Modern data privacy and security challenges
- AI and machine-learning security risks
- Privacy-preserving approaches
- Responsible AI
- Security defense and mitigation
Hands-on exposure
- Direct prompt injection
- Indirect prompt injection
- Security/privacy attack exercises
- Defensive mitigation exercises
- Privilege separation
- Spotlighting / separating data from instructions
- Blocklist limitations
Privacy-preserving ML exposure
- Differential privacy
- k-anonymity
- l-diversity
- Homomorphic encryption
- Secure multi-party computation
- Federated learning privacy considerations
Assessment
- Hands-on security/privacy attack activity -- 25%
- Defenses and mitigation activity -- 25%
- Literature-review presentation and peer evaluation -- 50%
Presentation topic: Federated learning in healthcare AI and the privacy question created by potentially sensitive information leaking through gradients, model updates, memorization, or inference.
Skills
AI and Machine Learning Security Risks, Responsible AI, Research and Critical Thinking, Cybersecurity Attack and Defence Strategies, Presentation and Peer Evaluation, Data Privacy and Security