日本語
2026-10-20 Evening edition
Evening editionAI News Daily
2026-10-20
AI agents are now active on both the attack and defense sides, making governance a pressing need.
2026-10-20 Evening edition
Today's Highlights
DomesticDomestic breach confirmed to use Chinese AI agent software "ARTEX"
2026-10-20 13:42
Feature UpdateMicrosoft unveils agent-focused security tools "Project Perception" and "Guardian Agents"
2026-10-19 22:00
2026-10-20 Evening edition
Domestic Breach Confirmed to Use Chinese AI Agent Software "ARTEX"
Published: 2026-10-20 13:42
- A GMO Internet Group-affiliated security firm said it confirmed that intrusions targeting multiple Japanese companies across industries used "ARTEX," AI agent software developed by a Chinese engineer.
- Chief Cabinet Secretary Kihara referenced the finding at an Oct. 20 press briefing, saying the government will share information with investigators and work to clarify the background.
Why it matters: With attackers now confirmed to use AI agents to automate much of the intrusion process domestically, companies face urgent pressure to strengthen AI-driven monitoring and detection on their own defense.
2026-10-20 Evening edition
Microsoft Unveils Agent-Focused Security Tools "Project Perception" and "Guardian Agents"
Published: 2026-10-19 22:00
- Microsoft's official blog warned that AI agents inside a company can become risks as an "agentic insider" even without being hijacked.
- It introduced "Project Perception," which discovers agents and assigns them identities and permissions.
- It also introduced "Guardian Agents," which continuously monitors risk and responds dynamically.
Why it matters: For companies rolling out AI agents at scale, traditional human-centric access management can't keep pace with these new risks, signaling that agent governance is becoming a must-have.
KEY CONTEXTProject Perception discovers agents and assigns identity/permissions. Guardian Agents continuously monitors risk and responds dynamically.
2026-10-20 Evening edition
Trend Overview
- AI agents are moving into practical use not just for productivity but as attack tools too.
- Defenders, including Microsoft, are rapidly rolling out security management features built specifically for agents.
2026-10-20 Evening edition
Summary
3 Key Takeaways
- A domestic breach was confirmed to use the Chinese AI agent software "ARTEX."
- The Chief Cabinet Secretary addressed it at a briefing, pledging investigative cooperation and further clarification.
- Microsoft announced new identity-management and monitoring features built for AI agents.
What to Watch Next
- Progress of the ARTEX investigation and its background
- Rollout scope and timing of Project Perception and Guardian Agents
- Whether other vendors follow with agent-focused security offerings