AI Security and Robustness

1.5 credits · College of Connected Computing - Professional

Security considerations specific to AI systems in production. Adversarial attacks and defenses; prompt injection and manipulation; red teaming methodologies for AI systems; reliability engineering for AI deployments; monitoring for model drift and degradation; incident response for AI failures; security considerations in multi-agent architectures. Students conduct red team exercises on deployed AI systems. Synchronous online with asynchronous components. [1.5]