AI Security Guide, Course

• Teach AI security research through studying attacks on and defenses for AI systems.
• Guide learners through five sequential sections on adversarial examples, jailbreaking, model tampering, and data poisoning.
• Provide access to webpages with topic overviews and hands-on coding exercises runnable in Google Colab or locally.
• Help learners identify vulnerabilities and expose weaknesses in current AI safety techniques.

This role equips learners with critical skills to identify and mitigate AI security risks, directly contributing to the safety and robustness of AI systems and reducing existential risks from advanced AI.

• Expertise in AI security, including adversarial examples, jailbreaking, model tampering, and data poisoning.
• Experience developing and delivering educational content or courses.
• Proficiency in Python and familiarity with Google Colab or similar environments.
• Strong understanding of AI safety research and current vulnerabilities.

Compensation and benefits are not specified in the posting.

The University of Chicago's Existential Risk Laboratory (XRL) conducts research and education to address existential risks, with a focus on AI safety and security.