Artificial Intelligence/Operations Research Workshop 2 Report Out

Trustworthy AI requires a multi-faceted approach: The workshop identified four foundational pillars—fairness, explainability, robustness/privacy, and human alignment—that must all be addressed together in AI system design rather than in isolation.

Fairness and explainability are interconnected challenges: The sessions on fairness and explainable AI/causality suggest these elements work together; understanding why an AI system makes decisions (explainability) is essential for identifying and correcting unfair outcomes.

Privacy and system robustness are critical safeguards: The robustness/privacy session highlighted that trustworthy AI systems must be both resistant to attacks or failures and protective of user data, indicating these are non-negotiable technical requirements.

Human oversight remains essential: The human alignment and human-computer interaction session emphasized that trustworthy AI cannot be achieved through technology alone—systems must be designed to work effectively with human judgment and maintain meaningful human control.

Solving real-world problems requires collaboration across disciplines: The workshop identified that addressing societal challenges through AI requires researchers from both artificial intelligence and operations research to work together, suggesting that responsible AI implementation benefits from diverse technical expertise.