AI Safety, Security, and Privacy

This unit examines the risks, harms, and legal/ethical implications that arise when AI systems are compromised and explores technical and organizational practices to reduce those risks. Students will study safety and alignment concepts, privacy principles and regulations, and accountability mechanisms to prepare AI systems for responsible deployment.

Learning Objectives

• Analyze likely threat scenarios and societal, organizational, and technical consequences of compromised AI systems, including harms to safety, privacy, and trust
• Apply leading security and technical safety practices (e.g., access controls, adversarial testing, monitoring, and fail-safe design) to develop mitigation plans for AI systems
• Evaluate privacy risks and applicable laws/regulations and recommend data-handling strategies (consent, minimization, de-identification limits, and governance) to protect personal information in AI applications
• Demonstrate how to establish accountability by defining roles and responsibilities, documenting decision and audit trails, and creating incident response and communication plans for AI failures