Sakana AI launches a research group for AI security and defense

Sakana AI announced its Security Research Group on November 20 to study how to secure AI systems and how organizations might use AI to coordinate cyber defense. The Tokyo-based team is led by Akihiro Umegai and David Ha. Sakana says the group will operate separately from its commercial business.
One priority is to prevent classes of software flaws during design and development rather than rely entirely on fixes after release. The team also wants to study whether organizations can share defensive signals reliably, with clear rules about what can be shared and who is accountable. A third focus is the defenders’ own AI agents: how to limit their permissions and make their decisions traceable.
Sakana says it plans to publish research, reproducible benchmarks and defensive guidance while withholding actionable exploit material. The group is hiring in Tokyo.
The announcement describes a research agenda, not a deployed defense network. Its value will depend in part on whether organizations can verify the group’s findings and use them without giving automated systems more authority than they need.
Sources and citations
Revision history
Published .