Microsoft launches AI cybersecurity platform developed in Israel

Microsoft has launched Project Perception, a cybersecurity platform powered by autonomous AI agents, and introduced its first dedicated AI model for cybersecurity, MAI Cyber 1 Flash. Developed in part at Microsoft's R&D center in Israel, the platform aims to shift cybersecurity from reactive to proactive defense. The MDASH system, first presented at Build, will be the first to incorporate the new AI model, detecting and analyzing software vulnerabilities. The launch positions Microsoft against competitors like Palo Alto Networks and CrowdStrike in the race to develop AI-agent-based security systems.

Microsoft launched Project Perception, a new cybersecurity platform built around autonomous AI agents, and unveiled MAI Cyber 1 Flash, its first dedicated AI model for cybersecurity missions. The company says the launch comes amid increasing use of AI by attackers to develop high-speed and large-scale attacks. Instead of adding AI to existing security products, Microsoft built a new architecture where AI agents are central to the defense system. The agents operate continuously, collecting data from enterprise sources, locating vulnerabilities, assessing risk, and helping address them before exploitation. The goal is to move from reactive defense to proactive risk reduction. Microsoft's R&D center in Israel played a central role: the MDASH system, developed in Israel and first presented at the Build conference last month, will be the first to incorporate MAI Cyber 1 Flash. MDASH detects and analyzes software vulnerabilities using AI. Microsoft says a significant part of the new platform's capabilities were developed in Israel. The move is strategic, as the cybersecurity market races to develop AI-agent-based systems. Microsoft aims to strengthen its position against competitors like Palo Alto Networks, CrowdStrike, SentinelOne, and Cisco, which are also investing in similar capabilities.