AIToday
Large Language ModelsRoboticsAI Safety & AlignmentTop Companies' AI MovesTop Companies AIPublished: Nov 2, 2026, 06:30 JST

NTT Docomo builds Physical AI連携基盤 to link AI agents and robots

Send on LINE
NTT Docomo builds Physical AI連携基盤 to link AI agents and robots

3 Key Points

  1. What happened

    NTT Docomo announced on Oct 27, 2026 that it developed a Physical AI Coordination Platform linking personal AI agents to multiple robots via the A2A protocol, with its own permission control, action-history tracking and context-aware behavior generation.

  2. Why it matters

    The platform automatically blocks disallowed robot actions before execution by checking a user's age and other attributes against venue-specific rules, and it avoids actions that make people uneasy—Docomo says such safeguards are needed before robots and AI agents can share human spaces.

  3. What to watch

    A demonstration of the platform is scheduled at NTT R&D FORUM 2026, held Nov 18–20 and Nov 24–25, 2026, at the NTT Musashino R&D Center.

WHO IT HITSRobot makers and food-service operators exploring service robots are the near-term audiences, since Docomo plans joint research with robot manufacturers and trials with various businesses. Consumers interacting with robots via a personal agent would be the end users the permission checks are designed to protect.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Docomo frames the problem as three gaps: robots each use different connection methods, there is no mechanism to judge and stop actions a user or setting should not allow, and robot motion can make people anxious or uncomfortable. The platform answers these by standardizing the connection on the existing A2A protocol and adding three functions Docomo developed itself—permission control, action-history tracking, and behavior generation tuned to surrounding conditions. Notably, user attribute data stays inside the platform, and only motion instructions are sent to the robot.

The company positions this as groundwork for robots that act automatically and appropriately for the person and situation. Its stated next steps—internal rollout, joint research with robot manufacturers, and trials with businesses—suggest the announcement is a first step toward validating real-world use rather than a product launch.

FAQ
How does the platform decide whether a robot may act?
An A2A Gateway checks the user's age and other attribute data, managed inside the platform, against rules registered for each usage setting, then stops any instruction that is not permitted before sending it to the robot.
What example use case does Docomo give?
In a restaurant, a user orders through a personal agent and the platform selects and operates a suitable serving robot, automatically declining to serve alcohol to minors or an allergen-containing food to allergic users, and routing around people in its path.
What are Docomo's next steps?
Docomo says it will verify practicality through internal deployment, joint research with robot makers, and proof-of-concept trials with various businesses.
Top Companies AIRead Original Article
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 articleRapidus-backed LSTC to build medical imaging system with 2nm chips