Fujitsu Unveils Multi-AI Agent Security Tech to Combat Emerging Cyber Threats
Image Credit: Ilya Pavlov | Splash
Fujitsu has announced the development of a cutting-edge multi-AI agent security technology, positioning it as the cornerstone of its AI service, Fujitsu Kozuchi. This innovative system leverages the coordination of multiple specialized AI agents to simulate cyberattacks, devise protection strategies, and ensure business continuity, thereby enabling companies and public organizations to proactively counteract emerging threats and vulnerabilities.
Comprehensive Technological Framework
Fujitsu’s multi-AI agent security system is built upon three foundational technologies:
1. Security AI Agent
At the heart of the system are three specialized AI agents designed for proactive security measures. Developed in collaboration with Ben-Gurion University of the Negev, a global leader in cybersecurity research, these agents include:
Attack AI Agent: Utilizes a tactics, techniques, and procedures (TTP) analogy engine to create effective attack scenarios targeting specific systems against new threats.
Defense AI Agent: Develops countermeasures tailored to a company's unique risk profile.
Test AI Agent: Constructs a virtual verification environment, or cyber twin, replicating the production system to assess the impact of proposed countermeasures.
The Test AI Agent conducts reciprocal simulations of attacks and defenses within the cyber twin environment, enabling the system to anticipate and mitigate new vulnerabilities through co-creative learning. This approach fosters collaborative and adversarial learning among AI agents with diverse expertise, addressing complex and unforeseen security challenges.
2. Multi-AI Agent Collaboration
This technology facilitates seamless collaboration between various AI agents, extending its applicability beyond security applications. A significant challenge in multi-agent collaboration is maintaining data confidentiality, privacy, and adherence to organizational policies. Fujitsu addresses this by integrating a secure agent gateway coordination function within each organization’s AI infrastructure. This feature allows for autonomous management of collaboration policies, ensuring smooth AI agent interactions without requiring explicit intervention from developers.
3. Generative AI Security Enhancement
Anticipating the rising threats to generative AI, Fujitsu, in partnership with Ben-Gurion University, has developed an LLM (Large Language Model) vulnerability scanner. This tool automatically and comprehensively assesses security resilience, identifying over 3,500 known vulnerabilities in generative AI systems, including a proprietary generative code vulnerability checker. Additionally, the system employs adaptive prompting technology to optimize attack prompt selection based on LLM responses, enhancing the accuracy of attack assessments. The LLM guardrails automatically enforce protective measures by detecting and rejecting malicious prompts, ensuring the secure operation of generative AI within enterprise IT systems.
[Read More: Microsoft Launches Zero Day Quest: A $4M Hackathon for AI and Cloud Security]
Strategic Partnerships and Field Trials
Fujitsu is set to begin field trials of its generative AI security enhancement technology in collaboration with Cohere Inc. starting December 2024. These trials will evaluate the effectiveness of the new security measures in real-world settings.
In January 2025, Fujitsu plans to release a portion of its multi-AI agent collaboration technology as open-source software (OSS) on OpenHands, an AI agent platform developed with significant support from Carnegie Mellon University. This move aims to foster broader adoption and innovation within the AI research community.
By March 2025, Fujitsu intends to launch the complete suite of its multi-AI agent security technologies globally, offering them on a trial basis to organizations worldwide.
[Read More: Cohere: The Canadian AI Unicorn Making Waves in the Industry]
Expert Endorsements
Associate Professor Graham Neubig of Carnegie Mellon University commented: “The development of AI agents is rapidly advancing worldwide. OpenHands, an open-source project strongly supported by Carnegie Mellon University, facilitates the development of diverse AI agent services, functionalities, and technologies by incorporating external expertise. Combining Fujitsu's multi-AI agent collaboration technology with OpenHands could significantly accelerate the development of cross-organizational multi-AI agent applications, which would be a remarkable achievement”.
Professor Yuval Elovici of Ben-Gurion University of the Negev added: “Fujitsu's innovative 'Cyber Twin' technology empowers companies to validate security policy compliance while simulating new attack and defense scenarios in a risk-free virtual environment. By enabling rigorous testing of business continuity without affecting real-world systems, this technology provides organizations with a powerful tool to swiftly address emerging new vulnerabilities and implement proactive, adaptive security strategies.”.
Seraphina Goldfarb-Tarrant, Head of Safety at Cohere Inc., stated: “Secure systems are mission-critical for enterprises to adopt and benefit from AI development. We're excited that our ongoing work with Fujitsu enables Japanese companies to leverage AI with our best-in-class multilingual models and security enhancement technology”.
[Read More: Unlocking $11 Million in Savings: The Power of Digital Twins in Business]
Future Directions
Fujitsu aims to expand its AI agent capabilities to automate a broader spectrum of security tasks. Future developments will include agents dedicated to secure design, incident response, and security audits. Additionally, Fujitsu will enhance its generative AI security measures to address vulnerabilities and mitigate hallucinations specific to Retrieval Augmented Generation (RAG) applications, thereby ensuring greater security and reliability for generative AI deployments.
[Read More: How Retrieval-Augmented Generation Revolutionizes AI Text Creation]
About Fujitsu
Fujitsu’s mission is to create a more sustainable world by fostering trust in society through innovation. As a leading digital transformation partner for clients in over 100 countries, Fujitsu employs 124,000 professionals dedicated to solving some of humanity’s most pressing challenges. The company’s diverse range of services and solutions integrates five key technologies—Computing, Networks, AI, Data & Security, and Converging Technologies—to drive sustainability transformation. For the fiscal year ending March 31, 2024, Fujitsu Limited (TSE:6702) reported consolidated revenues of 3.7 trillion yen (approximately US$26 billion), maintaining its position as Japan’s top digital services provider by market share.
[Read More: AI Explodes Data Growth, Tripling Since 2019: How to Balance Efficiency and Accuracy?]
License This Article
Source: Fujitsu
