As highlighted in the Summary of the 2018 Department of Defense (DoD) AI Strategy, the DoD is committed to leading in military ethics and AI safety. As a focal point of the strategy, the DoD seeks to ensure that developed and deployed AI systems are resilient (i.e., secure and robust). This webinar will provide an overview of adversarial AI, which are countermeasures that adversaries may deploy against our AI systems, and the evaluation steps and defenses needed to safeguard performance. Specifically, the following subtopics will be introduced:Â spanning adversarial AI countermeasures, adversarial AI access paradigms, adversarial AI attacks, system-level and algorithmic counter-AI defenses, counter-AI analysis, and demonstration of an open-source adversarial robustness toolbox. Overall, testing, evaluating, and assessing AI systems are important in identifying vulnerabilities and design recommendations to improve system resiliency.