#Japan #AI Governance #Osaka #RAYVEN #Next Innovator

RAYVEN's Remarkable Achievement at Next Innovator DEMO DAY

The world of artificial intelligence (AI) continues to evolve, and at the forefront of this transformation is RAYVEN, a company based in Osaka, Japan. Under the leadership of its CEO, Yoshihiro Suzuyama, RAYVEN has achieved a significant milestone by winning the top prize at the "Shido Next Innovator" DEMO DAY, a prestigious entrepreneur development program supported by the Ministry of Economy, Trade, and Industry and JETRO. This program culminates after an intensive six-month journey, where selected entrepreneurs refine their business ideas before pitching them, with only ten out of a hundred candidates advancing to the final stage.

The Winning Idea: Controlling AI Agents—A Necessity for Security

The core of Suzuyama's project revolves around ensuring that companies can manage the permissions and actions of AI agents effectively. With AI integration becoming more crucial in various industries, the program highlighted a pressing issue: many companies find their implementation halted at the security review stage. This stalemate occurs not only within finance, manufacturing, and public sectors but also in general businesses that are beginning to adopt AI technologies.

Questions regarding access limitations and accountability in the event of an incident are pivotal. These uncertainties often cause substantial delays in AI projects. Moreover, the traditional Model Context Protocol (MCP), used as a standard for connecting AI agents to internal systems and external tools, has been identified as a potential vulnerability that poses significant risks to businesses.

Risks Associated with Current Protocols

Recent studies suggest alarming trends in the stability and security of MCP implementations. For instance, a study from Equixly in March 2025 revealed that 43% of major open-source MCP servers were found to have command injection vulnerabilities. Among the prevalent issues are:

• - Tool Poisoning: This can happen when malicious definitions result in AI performing unauthorized actions, like altering or deleting customer data.
• - Privilege Escalation: This allows AI to access confidential information that it typically should not, potentially leading to data breaches.
• - Absence of Action Logs: Not having records of who instructed what to the AI could hinder incident recovery efforts and the auditing process.

Given these challenges, companies integrating AI into their operations must prioritize governance mechanisms.

Suzuyama's Vision for AI Governance

In the face of these challenges, Suzuyama has proposed a groundbreaking solution likened to an "ID card for AI agents". This aligns their operations with the established frameworks for human employees that include defined access levels, action logs, and accountability measures. The goal is to create a governance foundation that equates the operational capabilities of AI agents to those of human staff in terms of authentication, authorization, and auditing.

RAYVEN's technical initiatives in this domain focus on several key aspects:
1. Enforcement of Permissions: Clear definitions of permissible actions for AI, which help prevent tool poisoning and privilege escalation by differentiating between what actions are allowed and disallowed.
2. Comprehensive Record Keeping and Auditing: Every action by the AI is logged, enabling precise monitoring of who acted, when, and through which AI clients—critical in ensuring compliance and addressing incidents.
3. Data Protection through Split-Plane Composition: Sensitive data such as personal information remains secured within the client environment while only providing the management interface via cloud services.
4. Automated PII Masking: The governance systems proactively detect and mask personally identifiable information automatically, minimizing human error risks linked to data breaches.

Insights into the "Shido Next Innovator" Program

The "Shido Next Innovator" initiative, initiated in 2015, has played a critical role in shaping entrepreneurs in Japan. Sponsored by organizations such as Google and with support from the Ministry of Economy, Trade, and Industry, this program has nurtured over 1,150 innovators over the past decade. Participants undergo a rigorous six-month development period, out of which only the most outstanding ten finalists are chosen for the DEMO DAY pitch.

Suzuyama’s recognition is not just a personal victory but an acknowledgment of the broader potential of AI governance market development.

Looking Ahead

Buoyed by this award, Suzuyama is poised to accelerate the enterprise implementation of RAYVEN's governance foundation, targeting deployments in multiple industries by 2026. The strategic network built through the Silicon Valley connection and patented technology seeks to venture into the global market, promoting a secure environment for companies utilizing AI agents.

In his own words, "We stand on the brink of a new era where AI agents autonomously execute tasks. However, without clarity in accountability, businesses face risks regardless of their size or industry. Our groundwork aims to mitigate those risks before incidents occur, ensuring a reliable landscape for AI engagement across corporate sectors."

Invitation for Enterprises Considering AI Adoption

For businesses contemplating the integration of AI agents and facing challenges in security and operational management, RAYVEN offers comprehensive governance solutions that address permission management, action auditing, and policy enforcement all in one package.

• - RAYVEN is currently seeking three partner companies for an exclusive, no-cost proof of concept (PoC) to test their integration in a real-world environment.
• - Comprehensive enterprise explanation materials detailing technical overviews and implementation processes are available.
• - Customized demo requests are also welcomed to align with specific business environments.
• - For inquiries, contact: [email protected], or schedule a free consultation via this link.

RAYVEN stands ready to assist in paving the path toward secure and effective AI deployment in the business world.