Case Study: Microsoft achieves standardized AI security testing for 70+ products with NetSPI

How Microsoft and NetSPI Partnered to Build a Standardized AI Security Framework Securing 70+ Products

Microsoft partnered with NetSPI to address a major challenge in validating the security of its AI-powered products. As Microsoft expanded its AI offerings, it needed a structured, repeatable way to assess emerging AI vulnerabilities, establish scope and severity, and ensure its solutions were not only secure but also trustworthy and safe for public use.

NetSPI co-developed the AI/ML Security Assessment Framework with Microsoft, combining existing work, expert review, and real-world testing to create a measurable standard for AI security assessments. The framework has been used in 28 tests across more than 70 products, uncovering 126 vulnerabilities and giving Microsoft clearer evidence of testing, remediation, and overall AI security posture.

Microsoft

Daniel Moore

Principal Security Assurance Engineer