University of Tennessee, Knoxville

A successful first Global Dialogue on AI Governance should move beyond high-level principles and produce concrete, operational alignment across stakeholders. In my view, three outcomes would define success. First, the Dialogue should establish a shared framework for translating AI capabilities into governance-relevant constructs, particularly around safety, accountability, and human oversight. This includes agreement on how to represent AI system intent, uncertainty, and risk in ways that are both technically grounded and interpretable to policymakers. Second, it should produce actionable pathways for integrating governance into system design, not just post hoc regulation. This means embedding auditability, traceability, and constraint-awareness directly into AI pipelines. Third, the Dialogue should catalyze cross-sector testbeds and benchmarks that bridge policy and practice. Ultimately, success would mean shifting AI governance from reactive oversight to proactive, system-integrated design, grounded in technical realism, validated through experimentation, and adaptable to rapidly evolving AI capabilities.

Transparency, accountability, and human oversight;Safe, secure and trustworthy AI

My selection of Safe, secure and trustworthy AI and Transparency, accountability, and human oversight reflects a focus on ensuring that AI systems operating in real-world, safety-critical environments remain reliable and aligned with human intent. In domains such as robotics and cyber-physical systems, AI directly interacts with physical environments and human collaborators. This makes safety and trustworthiness a design requirement rather than a high-level goal. Systems must incorporate constraint-aware decision-making, real-time monitoring, and adaptive control to manage uncertainty and risk. At the same time, transparency and accountability are essential for effective governance. AI systems should provide interpretable representations of their decisions, confidence, and constraints, enabling human operators to understand, supervise, and intervene when necessary. This includes traceable decision pipelines and mechanisms for continuous oversight. These priorities support a shift from reactive regulation to proactive, system-integrated governance, where AI is designed to be auditable, controllable, and robust in dynamic environments.