Governance

The Problem

Technical safety is only half the picture, a well-behaved model can still cause harm without the right rules and oversight. Regulation lags behind technology, existing legal frameworks weren't built for AI-augmented operations, and synthetic media raises urgent questions about provenance and consent as AI development crosses borders faster than shared standards can form.

What We're Working On

• AI and international law — How AI-augmented cyberattacks challenge existing humanitarian law frameworks.
• Synthetic media governance — Detection and localization approaches to mitigate risks from generative models.
• Field-building — Growing the AI safety and governance community, with a focus on Japan.