Austin Wyatt

AI & National Security for Middle Powers

AI safety conversations are often framed around the choices of frontier developers and great powers. This talk examines a different vantage point: Considering how middle powers can responsibly adopt AI-enabled systems into national security functions, without necessarily controlling the underlying models, data or compute. In this talk, I argue that responsible AI principles function as a strategic enabler rather than a constraint for states such as Australia. Working in coalition with like-minded partners, middle powers can leverage combined market power and normative credibility to shape the behaviors of upstream developers, advance practical standards and guardrails, and limit escalation risk in our region.