Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Course Outline
Introduction to AI Red Teaming for Government
- Understanding the AI threat landscape within government systems
- Roles of red teams in enhancing AI security for government applications
- Ethical and legal considerations for government agencies
Adversarial Machine Learning for Government
- Types of attacks: evasion, poisoning, extraction, inference in the context of government systems
- Generating adversarial examples (e.g., FGSM, PGD) to test government AI models
- Targeted vs untargeted attacks and success metrics for government applications
Testing Model Robustness for Government
- Evaluating robustness under perturbations in government AI models
- Exploring model blind spots and failure modes specific to government use cases
- Stress testing classification, vision, and NLP models for government applications
Red Teaming AI Pipelines for Government
- Attack surface of AI pipelines: data, model, deployment in the government context
- Exploiting insecure model APIs and endpoints within government systems
- Reverse engineering model behavior and outputs for government applications
Simulation and Tooling for Government
- Using the Adversarial Robustness Toolbox (ART) for government AI projects
- Red teaming with tools like TextAttack and IBM ART in government settings
- Sandboxing, monitoring, and observability tools tailored for government use
AI Red Team Strategy and Defense Collaboration for Government
- Developing red team exercises and goals aligned with government objectives
- Communicating findings to blue teams within government agencies
- Integrating red teaming into AI risk management for government systems
Summary and Next Steps for Government
Requirements
- An understanding of machine learning and deep learning architectures for government applications.
- Experience with Python and ML frameworks (e.g., TensorFlow, PyTorch) in a public sector context.
- Familiarity with cybersecurity concepts or offensive security techniques for government operations.
Audience
- Security researchers for government agencies.
- Offensive security teams within the public sector.
- AI assurance and red team professionals for government organizations.
14 Hours
