BREAKING
Japan Drafts First Revision of AI Plan
Review of AI Law for Frontier Risks
Claude Mythos Benchmark Scores
SWE-bench93.9
USAMO97.6
CyberGym83.1
What the Draft Sets Out
1Collect risk info
2Strengthen AISI
3Evaluate models
4Emergency response
Promotion Law vs Penalty Debate
Current AI Law2025
Promotion-oriented
No penalties
Advice and guidance
Proposed Review
Penalties for misuse
Emergency response
Stronger oversight
Cabinet Decision Targeted for July
AI NEWS BLITZ
Japan has drafted the first revision of its national AI Basic Plan.