AIToday
Large Language ModelsAI Regulation & PolicyAINOWPublished: Nov 4, 2026, 01:00 JST

Japan's AI事業者ガイドライン(第1.2版) targets agent governance

Send on LINE
Japan's AI事業者ガイドライン(第1.2版) targets agent governance

3 Key Points

  1. What happened

    Japan's Internal Affairs and Communications Ministry and the Ministry of Economy, Trade and Industry published the AI事業者ガイドライン(第1.2版) on 2026年3月31日, adding AI agent definitions and risks like unintended orders, file deletion, and external attacks.

  2. Why it matters

    The guidelines are not legally binding, but trading partners and audit firms reference them when checking a company's AI management setup, so firms may be asked to show agent-level controls.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

The guidelines arrive as agent-building tools have become common in ordinary departments. Microsoft Copilot Studio and Dify let business-side staff create agents, which makes it hard for IT to know who built an agent, what it connects to, and who maintains it. When a creator transfers or leaves, an agent with permissions can remain without an owner. That gap is the backdrop for the guidelines' focus on giving each agent its own ID and a named responsible person.

Alongside Japan's guidelines, other frameworks are moving in the same direction: Singapore's IMDA published Model AI Governance Framework for Agentic AI on 2026年1月22日, and the EU agreed in May 2026 to delay applying high-risk AI obligations. International standards and vendor tooling include ISO/IEC 42001, NIST AI RMF, and agent management tools from Microsoft, Amazon, and Google. For Japanese companies, the practical link is that public guidelines can serve as the stated basis when internal rules are reviewed by management or auditors.

FAQ
Is agent governance required by law in Japan?
No. The AI事業者ガイドライン(第1.2版) has no legal binding force, but it is referenced by trading partners and audit firms when they check a company's AI management system.
What is the difference between agent governance and traditional AI governance?
Traditional governance controls outputs such as accuracy and fairness, while agent governance controls actions and permissions. Agents can send emails and update data without a human reading each result, so existing generative AI rules need added permissions, approval, and stop rules.
What does Gartner predict about agentic AI projects?
Gartner predicts over 40% of agentic AI projects will be canceled by 2027年末 because of rising costs, unclear business value, and insufficient risk management.
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleCoworkOS claims parity with Codex, DeepSeek Harness