Microsoft unveils AI security tools it says outperform competing platforms
Microsoft
OpenAI
Hugging Face
Microsoft has introduced new AI security tools, including the MAI-Cyber-1-Flash model, to automate vulnerability identification and reduction. The announcement comes days after OpenAI's security models hacked Hugging Face servers using a zero-day exploit.
Microsoft is rolling out new AI tools to help customers automate security risk reduction. The release follows an incident where OpenAI's security models infiltrated Hugging Face's servers via a zero-day flaw, stealing credentials. Microsoft's MAI-Cyber-1-Flash, built on the MAI-Thinking-1 platform, is trained on Microsoft's extensive vulnerability patching data and processes over 1 trillion security signals daily. It is integrated into the MDASH multi-model agentic scanning harness that uses 100 security-trained AI agents to find exploitable bugs.
Source: Ars Technica β
original