You are currently viewing Microsoft Develops Secure AI for US Intelligence Services

Microsoft Develops Secure AI for US Intelligence Services

Rate this post

Microsoft, a company heavily invested in artificial intelligence (AI), has created a new AI system specifically for United States intelligence agencies. Unlike other AI tools, like Microsoft’s Copilot, this one is designed to operate without needing to connect to the internet, making it more secure.

According to Bloomberg, this AI model is the first of its kind to be completely disconnected from the internet, a feature crucial for maintaining security. Most AI systems rely on internet connections to learn and analyze data, but Microsoft wanted to ensure the utmost security for the US intelligence community.

This new tool, based on Microsoft’s GPT-4 technology, has been in development for 18 months. It allows intelligence services to analyze classified data without the risk of leaks or hacks that could endanger national security.

William Chappell, Microsoft’s CTO for Strategic Missions and Technology, explained that the AI system can read, analyze, answer questions, and even write code without needing an internet connection. Importantly, it does not learn from the data it processes, ensuring that sensitive information remains secure.

Sheetal Patel, assistant director of the CIA for the Transnational and Technology Mission Center, highlighted the importance of having generative AI for intelligence purposes, emphasizing the need for security in such endeavors.

Source: techradar