Microsoft is integrating AI models, including Anthropic Mythos, into its security development lifecycle.
Techzine reported on April 23 that the integration focuses on having AI find vulnerabilities in the early stages of development rather than after software is released. Microsoft aims to significantly reduce the time from discovering vulnerabilities to resolving them.
Microsoft said it believes current AI models can find vulnerabilities on their own and combine multiple small flaws to create real attack scenarios. As threats emerge faster, it sees a need to review existing security strategies.
Microsoft made clear it is not relying on a single model. Since no model fully covers every aspect of cybersecurity, it chose to combine multiple AI models.
Vulnerabilities detected by AI will be handled through the existing process at the Microsoft Security Response Center.
Chi-gyu Hwang
delight@d-today.co.kr