AI Product Security: Incident Response

AI is experiencing a number of growing pains as it evolves, and these are too frequently escalating into high-profile incidents. In this course, instructor Malcolm Shore covers a wide range of issues which are relevant to AI applications, including the traditional cybersecurity issues as well as more contemporary and AI-specific issues associated with inadequately protected data, inadequately protected AI applications, untested models, and supply chain issues with training data and model construction. From prompt and thought injections to gaining operating system shells, this course provides a hands-on approach to understanding how you can respond to and minimize the damage caused by an incident.

Learning objectives
Analyze the scope and types of incidents that can affect AI applications and their potential impacts.
Identify events and indicators that may constitute emerging AI incidents through hands-on lab exercises.
Evaluate and execute appropriate incident response procedures for different types of AI incidents.
Formulate effective communication strategies for stakeholders during AI incidents.
Design preventive measures to protect AI applications based on common incident patterns and vulnerabilities.