Cornell Launches AI Security Initiative with Amazon Support

Cornell University has partnered with Amazon to establish the AI4AI initiative aimed at enhancing the security of agentic AI systems. The initiative will focus on developing safety protocols to prevent AI agents from generating harmful code. Principal investigators Alexandra Silva and Vitaly Shmatikov will lead the project, which includes faculty from various departments. The initiative addresses concerns that AI agents can produce incorrect or malicious code that hackers could exploit. The collaboration aims to create a security framework with rules and verification checks to ensure the integrity of AI-generated software. This initiative is particularly relevant as AI technology continues to evolve rapidly, raising security challenges. Amazon's support is seen as a significant step towards addressing these pressing issues in AI development.

Key Points: • Cornell and Amazon have launched the AI4AI initiative to improve AI security. • The initiative aims to develop protocols to prevent AI from generating harmful code. • Principal investigators include notable Cornell faculty specializing in AI and security.