In-Depth Discussion on Cyber Defense Strategies in the AI Era! AI Data Inc. Reports on 'AI Agent × AX Forum ~Cybersecurity~'

AI Data Inc., a company specializing in enterprise data and AI utilization (Headquarters: Minato-ku, Tokyo; President and CEO: Takahito Sasaki; hereinafter referred to as AI Data), held the AI Agent × AX Forum on May 21, focusing on AI applications and AX in the field of cybersecurity—one of the 17 strategically prioritized sectors by the Japanese government.

With the widespread adoption of generative AI, cyberattacks are rapidly becoming more sophisticated, demanding new security strategies from enterprises. This forum facilitated discussions based on cutting-edge insights in areas such as AI advisors, AI trust, zero trust, and digital forensics.

For detailed event report, please visit:

■ Session 1

"Cybersecurity is not an IT issue — A two-layer strategy of management decisions and execution capabilities"

Takahito Sasaki, Representative, AOS Group

Mr. Sasaki emphasized that cybersecurity should not be viewed solely as an IT department issue, but as a strategic business imperative to protect the entire organization. He pointed out that recent cyberattacks go beyond data breaches and system outages, significantly impacting business continuity and corporate value. As AI accelerates the speed and sophistication of attacks, he explained the importance of a two-layer strategy: a management decision layer determining what to protect, and an operational execution layer responsible for continuous defense. He further stressed the importance of leveraging AI to bridge management and frontline operations, enabling organizations to recover swiftly even after an attack.

■ Session 2

"Cybersecurity in the Age of AI Agents — Emerging Threats and Countermeasures Targeting Organizations, Insights from Automotive and Robotics Research"

Seiki Hara, Executive Officer and Head of Technology, VicOne Inc.

Mr. Hara presented the latest trends in cybersecurity within physical AI domains such as automobiles and robotics. He highlighted attacks targeting AI systems themselves, misuse of AI agents, and unauthorized intrusions into robots, illustrating how AI proliferation introduces new risks. He analyzed that advancements in generative AI are accelerating vulnerability discovery and attack code generation, further tilting the balance in favor of attackers. On the other hand, he stated that defenders can also leverage AI to enhance threat detection and analysis, asserting that we have entered an era where "AI-powered attacks must be countered with AI-powered defense."

■ Session 3

"Secured by AI Agents — Real-time Demonstration of the Two-Layer Strategy with AI InfraSec on IDX"

Daisuke Shida, Director and CTO, AI Data Inc.

Mr. Shida introduced AI Data's approach to supporting cybersecurity through two layers: management decision-making and frontline execution. During his presentation, he demonstrated "AI InfraSec on IDX," integrating patents, technical information, and access logs, and showcasing how "seven AI advisors" perform threat analysis and risk assessment. The system demonstrated end-to-end support during incidents—from assessing damage to formulating countermeasures and supporting executive decision-making. He advocated for treating cybersecurity not merely as defense, but as a strategic foundation supporting corporate competitiveness and business continuity.

■ Session 4

"Where should your organization start? — Design and Practice of AI Trust"

Yoji Watanabe, Representative Director and CTO, Cybersecurity Cloud Inc.

Mr. Watanabe pointed out that while generative AI adoption is accelerating, governance and security measures are lagging. He emphasized the importance of organizational operations centered on "AI Trust." With the rise of AI agents, new risks such as data leaks and privilege management have emerged. He introduced five key principles: "Trust by Design," "Least Privilege," "Separation of Data and Intelligence," "Observability," and "Human Override," presenting a practical framework for trustworthy AI governance and data management. He concluded that instead of halting AI adoption, organizations should mature their operations through rule-setting and enhanced visibility.

■ Session 5

"We're Fully Using AI — How Attackers Are Automating Cyberattacks with Powerful AI"

Koji Morii, Representative Director and President, Leon Technology Inc.

Mr. Morii explained from an attacker's perspective how generative AI is transforming the cyberattack lifecycle. He noted that AI automates target reconnaissance and attack scenario creation, enabling attackers to launch larger-scale and more precise attacks than ever before. He particularly highlighted efficiency gains in the "reconnaissance" and "delivery" phases of the cyber kill chain. However, he emphasized that existing security measures and fundamental practices—such as asset management and log monitoring—remain effective. He concluded that defenders must also leverage AI to counter AI-powered attackers.

■ Session 6

"Invisible Technology of 'SA1' — Zero Trust for Legacy Medical and Industrial Equipment"

Hideto Takeuchi, Representative Director, Verification Technology Inc.

Mr. Takeuchi introduced the cybersecurity challenges faced by legacy medical and industrial equipment, and presented "SA1," an invisibility technology developed as a solution. He explained an innovative approach that makes devices "invisible, untouchable, and non-targetable"—achieving security without major modifications. By controlling communication paths and concealing the very existence of devices, SA1 enhances protection while preserving existing infrastructure. He presented this as a practical path toward zero trust implementation in critical environments such as healthcare and manufacturing, where system downtime is unacceptable.

■ Session 7

"Accelerating Initial Incident Response! Forensic Techniques to Instantly Confirm Personal Data and Operation Logs"

Yoshiaki Morita, Representative Director and CEO, Legal Data Inc.

Mr. Morita discussed the critical role of digital forensics in initial incident response. With increasing incidents of ransomware and unauthorized access leading to data breaches, organizations must rapidly and accurately assess the impact. He introduced forensic investigation methods that analyze logs and operation histories on PCs and servers to visualize the presence of personal data and access patterns. He explained that accurate fact-finding helps prevent further damage and accelerates executive decision-making, emphasizing the importance of log management and evidence preservation during normal operations.

■ Special Panel Session with Speakers

In the special session, all speakers gathered to discuss the future of cybersecurity in the AI era. Topics included the escalation of AI-powered attacks and defenses, risks associated with physical AI, and the growing importance of zero trust and forensics. The panelists agreed that as AI adoption deepens, the importance of data governance and management will continue to rise.