Governance

The Problem

Technical safety is only half of the picture. Even a well-behaved model can cause harm if it is deployed without the right rules, oversight, and accountability. Governance asks how societies, institutions, and laws should adapt to increasingly capable AI.

Key challenges include:

• Regulatory lag — Technology moves faster than the laws and norms meant to govern it, leaving gaps in accountability.
• Legal frameworks under strain — Existing frameworks, including international humanitarian law, were not designed for AI-augmented cyber operations or autonomous systems.
• Content governance — Synthetic media and generative models raise urgent questions about provenance, consent, and misuse.
• Global coordination — AI is developed and deployed across borders, making shared standards and community-building essential.

What We're Working On

• AI and international law — Examining how AI-augmented cyberattacks challenge existing international humanitarian law frameworks.
• Content and synthetic media governance — Developing detection and localization approaches to mitigate risks from increasingly realistic synthetic media.
• Responsible AI practices — Studying the institutional and policy mechanisms that keep AI development accountable.
• Field-building — Growing the AI safety and governance community, with a particular focus on Japan.