AIM Intelligence joins Microsoft AI security partner program

AIM Intelligence has been selected as an early partner in Microsoft’s MDASH AI security system, becoming the only publicly disclosed Korean participant. The partnership puts the Seoul-based firm alongside IBM, Accenture and PwC as Microsoft pushes to set security standards for autonomous AI agents and physical AI systems.

Why it matters: - AIM Intelligence’s selection gives the South Korean AI security specialist a role in shaping emerging standards for autonomous AI agents and physical AI systems. - The partnership places AIM Intelligence inside a global security ecosystem as companies race to secure agentic AI and other next-generation models. - Microsoft’s MDASH program could influence how software vulnerabilities are found and validated across AI-driven systems.

What happened: - AIM Intelligence said it was chosen as an early partner in Microsoft’s MDASH, or Multi-moDel Agentic Scanning Harness, in limited preview. - Among publicly disclosed participants, AIM Intelligence is the sole Korean company in the initiative. - Microsoft’s partner group also includes IBM, Accenture, PwC, Infosys and TCS. - AIM Intelligence said it will help with initial testing and practical evaluation of MDASH.

The details: - MDASH is a multi-model agentic AI vulnerability detection system that uses multiple specialized AI agents to identify and validate software vulnerabilities. - Microsoft’s system has posted an 88.45% score on the public CyberGym benchmark. - That result put MDASH at the top of the CyberGym leaderboard. - Microsoft said MDASH also outperformed Anthropic’s Mythos preview in independent evaluations. - AIM Intelligence develops vulnerability detection and mitigation tools for large language models, AI agents, multimodal models and physical AI systems. - The company uses automated AI red-teaming and real-time guardrail technology. - AIM Intelligence supplies AI security tools to customers in finance, manufacturing, telecommunications and the public sector.

Between the lines: - The partnership suggests Microsoft wants broader industry input as AI systems move from chatbots to agents that can act inside software and physical environments. - AIM Intelligence’s role may help the company test its red-teaming and guardrail products against newer classes of AI risk. - CEO Sangyoon Yu said traditional security approaches are no longer sufficient as AI agents and physical AI systems become more common.

What's next: - AIM Intelligence plans to expand its security response capabilities for threats that emerge when AI agents autonomously explore and operate within systems. - The company said the Microsoft partnership will help it advance automated red-teaming and real-time guardrails against next-generation AI risks. - The collaboration is expected to support further technical work inside Microsoft’s broader AI security network.

The bottom line: - AIM Intelligence’s inclusion in MDASH gives the Korean firm a visible role in a high-profile effort to define AI security for autonomous and physical systems.