Case Study: Microsoft builds trusted AI security testing with NetSPI

Microsoft secures 70+ products with NetSPI’s AI testing framework

Microsoft, a global leader in AI, faced the challenge of ensuring its AI solutions were secure, ethical, and trustworthy for public use. They needed a standardized framework to validate the resilience of their AI implementations against a new class of vulnerabilities, as their existing security assessments lacked a structured methodology for AI. They partnered with NetSPI to co-develop a solution.

NetSPI collaborated with Microsoft to build an AI/ML Security Assessment Framework from the ground up. This provided a structured, repeatable process for testing. The results were significant: NetSPI performed 28 tests using the framework, informing over 70 products and discovering 126 vulnerabilities. This standardized approach gave Microsoft a measurable method to evaluate its AI security posture and enhanced customer trust by providing clear evidence of their commitment to security.

Microsoft