Managing AI Agents in the Physical World: Control, Security & Safety
Managing AI Agents in the Physical World: Control, Security & Safety
Physical AI × Robotics × Enterprise AI
As AI agents become increasingly connected to robotics, autonomous systems, industrial infrastructure, and enterprise operations, a critical question emerges:
Who controls AI agents, what are they allowed to do, and how can we ensure accountability when things go wrong?
Join us for an interactive industry roundtable in Tokyo, bringing together founders, engineers, technical leaders, and enterprise practitioners from the Physical AI, robotics, semiconductor, and AI infrastructure ecosystems.
We'll explore the real-world challenges of deploying and managing AI agents safely, from access permissions and operational security to auditability and human oversight.
This is not a traditional product presentation. It's an opportunity to exchange practical experiences, discuss emerging challenges, and connect with people building and deploying AI systems in the real world.
What We'll Discuss
AI Agent Control & Permissions: Who should authorize an AI agent to access systems or execute actions?
Security & Risk Management: How can organizations limit high-risk operations and protect critical infrastructure?
Auditability & Human Oversight: How do we track AI agent activities, maintain accountability, and introduce human approvals?
Physical AI Meets Enterprise Infrastructure: What changes when AI agents interact with robotics, industrial systems, and operational workflows?
Enterprise Adoption: What technical and organizational challenges must be solved before autonomous AI systems can be widely deployed?
Featured Company: SOSI / 5xRuby
We're pleased to welcome SOSI, a privileged remote access and session auditing platform for both human operators and AI agents developed by Taiwan-based 5xRuby.
SOSI provides role-based permissions, rule-based interception of high-risk commands, full session recording, and ticket-based, time-limited access approvals, so every action is traceable.
During the event, Mu-Fan Teng (Ryudo) will join the panel discussion and present a 10-minute demonstration of SOSI's approach to managing AI agent access and operational security.
https://www.sosi.com.tw/en/
Event Agenda
November 30, 2026 | Tokyo | 6:00–8:00 PM JST
18:00 – Doors Open / Registration
18:25 – Welcome Remarks (Kaori + JETRO)
18:30 – Technical Panel / Roundtable
18:55 – Interactive Audience Q&A
19:15 – SOSI Live Demo
19:25 – Networking
20:00 – Event Closes
Speakers
Mu-Fan Teng (Ryudo), SOSI / 5xRuby
Physical AI / Robotics / Enterprise Guest, to be announced
Moderator, to be announced
Event Host / MC: Kaori Rei, AI Founders' Mindset / KaoriRei Inc.
Who Should Attend?
This event is designed for:
Physical AI and robotics founders, engineers, and technical leaders
Enterprise AI and infrastructure teams
Semiconductor, edge AI, and industrial technology professionals
AI security, governance, and platform engineering specialists
Developers building autonomous systems and AI agents
Investors and ecosystem leaders interested in real-world AI deployment
Whether you're building autonomous systems, managing enterprise AI infrastructure, or exploring how AI agents can operate safely in complex environments, we welcome your perspective.
Event Partners
Organizer:
Venue & Co-organizer: JETRO (Japan External Trade Organization), pending final confirmation
Registration
Attendance is free. Registration is required, and space is limited.
To encourage meaningful technical discussions and networking, registrations may be reviewed to maintain a relevant mix of participants.
We look forward to meeting you in Tokyo!