Google Discusses Expanding Fairwind AI Cyber Defense to Japan's Government and Critical Infrastructure
·2026.11.17 10:35
Key point
Google is in discussions to expand its Fairwind AI cyber defense program to Japan, following the country's allocation of ¥10 billion in emergency cybersecurity funds.
Details
- Google Cloud CEO Thomas Kurian visited Japan on November 16 to discuss cybersecurity strategy with government officials, with reports indicating potential expansion of the Fairwind Program to the country.
- The Fairwind Program, announced on September 2, 2026, provides early access to AI models for defenders at governments, healthcare providers, and telecommunications companies.
- The program offers the cybersecurity-focused 'Gemini for cyber' model and CodeMender, a code security agent that autonomously inspects and repairs code.
- Gemini 4 Carbon for cyber ranked first in benchmarks for vulnerability discovery (SEC-bench Pro), exploit validation (ExploitGym), and patching (CWE-bench v1).
- Specific terms for the Japan expansion, including participating agencies, target models, and launch dates, remain unconfirmed.
Fairwind Program Details
- The program restricts use to defensive and academic research, prohibiting redistribution, resale, and non-defensive malware analysis.
- Google states it works with thousands of partners globally under strict access controls and multifactor authentication requirements.
Japan Cybersecurity Context
- On November 6, the Japanese government adopted an emergency cybersecurity package allocating about ¥10 billion in reserve funds.
- Prime Minister Sanae Takaichi set a target for critical infrastructure operators to report attacks within a few hours if possible, and at least on the same day.
- The emergency measures call for earlier AI-assisted vulnerability assessments and stronger intelligence collection through cooperation with overseas technology companies.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.