Japan based Sakana AI announced setting up of Security Research Group, an independent research thinktank on AI driven cybersecurity with hirings sought for senior researchers.
According to a statement released by the AI company on Thursday, the Sakana AI Security Research Group will function independently of the Japanese Frontier AI Lab’s business activities and will focus on sharing research findings, setting up benchmarks and guidelines.
“As AI uncovers software flaws and security incidents increase, we will pursue research and development aimed at fundamentally improving security by treating it as a quality issue and preventing defects from being introduced during design and development. The group operates independently of our business activities, sharing research findings, reproducible benchmarks, and practical guidelines while collaborating with government, industry, and academia,” read a statement from Sakana AI.
Also Read:In Cybersecurity, AI is Both Lock and Lockpick: WEF & KPMG Report
The group will be working on two main areas: ensuring cyber defenses created through AI and securing various technologies that are used in these systems. Sakana AI argued that the idea behind its work is that by cooperating the society can build new strong systems instead of focusing on the previous way of building strong systems of particular individuals.
Sakana AI also announced vacancy for senior researchers in the newly formed Security Research Group.
“We will work on three questions: how to build quality in, so that whole classes of flaws are prevented and automated fixes can be verified; whether defense with AI agents can be coordinated across organizations at machine speed (an open, early-stage direction); and how to keep AI itself secure and safe, starting with the defenders’ own AI agents. We intend to share what we learn openly, through papers, reproducible benchmarks,” said Sakana AI.
The new research group will publish its findings through research papers, reproducible benchmarks and practical engineering guidelines, while collaborating with government agencies, critical infrastructure operators and the wider cybersecurity community. Sakana AI said it will approach the research cautiously given the dual-use nature of cybersecurity discoveries, which could potentially benefit both defenders and attackers.
Also Read: Japan’s Sakana AI Taps Ex-NSA Director and Ex-CIA Official for Defense and Cybersecurity Roles






