Google DeepMind's AI Agent CodeMender Autonomously Fixes Security Vulnerabilities in Software Code

Google DeepMind launches CodeMender, an AI agent that autonomously detects and fixes security vulnerabilities in software, having already submitted 72 security patches across 4.5 million lines of open-source code including fixes for exploits used in major attacks.

• Google DeepMind unveils CodeMender, an AI agent that autonomously detects, patches and rewrites vulnerable software code to prevent future exploits
• CodeMender has already submitted 72 security fixes to open-source projects spanning over 4.5 million lines of code, including applying safety annotations to the libwebp library exploited in a 2023 iOS attack
• The system combines Gemini Deep Think models with program analysis techniques and includes validation tools, though it remains in research phase with all patches requiring human review before submission