GenAI CBRNE Cyber Expert
Summary
Red-team generative AI systems to uncover vulnerabilities in cyber CBRNE threat modeling, adversarial prompts, and multi-turn safety evaluations.
- Conduct taxonomy audits and safety evaluations
- Design adversarial prompts and jailbreaks
- Develop cyber CBRNE attack scenarios
- Develop red team test suites and evaluation rubrics
- Evaluate model assistance for cyber enabled CBRNE threats
- Execute adversarial red teaming of generative AI
- Identify and document failure modes and attack patterns
- Maintain knowledge of emerging GenAI attacks and cyber threat intelligence
- Run multi turn scenario evaluations