Project Perception & AI-Cyber-1-Flash aim to proactively identify, assess, and mitigate cyber risks through independent AI agents and specialized threat-analysis.
Microsoft announced today Project Perception, a new AI-powered, agent-based security system, designed to help organizations continuously detect, assess, and reduce cyber security risks.
The company also introduced Microsoft AI-Cyber-1-Flash, its first AI model built specifically for cybersecurity, designed to analyze software vulnerabilities.
The speed and scale of cyberattacks continues to progress with AI, while security defenders must continue to rely on tools and methods to protect increasingly complex digital environments using approaches developed for a less complex digital environment. As a result security teams are often forced to gather signals, context, and risk information from enormous volumes of data, making it difficult to keep up with emerging threats.
To tackle this challenge, the company is launching a new approach that combines contextual security, specialized AI models, and autonomous agents to continuously identify, evaluate, and help mediate risks across an organization’s digital assets.
Key Highlights
Project Perception, a new agent-based security system designed to help organizations continuously detect, evaluate, and reduce security risks, allowing security teams to shift from a reactive response to proactive risk mitigation. Perception coordinates three types of specialized AI agents, Red, Blue and Green, that work together to continuously strengthen an organization’s overall posture over time. The Red agents work to identify potential attack vectors and vulnerabilities before they can be exploited, while the Blue agents investigate the findings, consider the security context, and determine what poses a significant risk. Lastly, the Green agents will work on the implementation of corrective actions as they strengthen defenses throughout the environment.
Microsoft also launched its first Artificial Intelligence model, Microsoft AI-Cyber-1-Flash, which was developed specifically for cybersecurity. It is designed to analyze software vulnerabilities and developed as part of the company’s broader multi-model strategy, which combines state-of-the-art AI with specialized models tailored to specific security tasks.
Drawing on Microsofts tensive security data, threat intelligence, and real-world operational experience, Project Perception is designed to help organizations move beyond reactive cybersecurity measures and toward a more adaptive and proactive approach to managing security risks.
