The Future of Espionage is Here—Microsoft’s AI Chatbot for Intelligence Work

Microsoft has recently introduced a new GPT-4-based generative AI model that is designed to be used specifically by US intelligence agencies. And the best thing is that this model will work without connecting to the internet.

This is the first time that Microsoft has taken the lead in deploying a major language model in a secure setting and is designed to allow spy agencies to analyze top-secret information without connectivity risks—and to allow secure conversations with a chatbot similar to ChatGPT and Microsoft Copilot.

According to a report, the new AI service can help intelligence agencies use generative AI to process classified data while avoiding the risks of data breaches and hacking.

Microsoft’s Chief Technology Officer, William Chappell, stated that developing this new system involved more than 18 months of work to modify an AI supercomputer in Iowa. The modified GPT-4 model is designed in such a way that it can read files provided by the users but cannot connect to the internet.

This new service was activated last Thursday and is now available to more than 10,000 individuals in the intelligence community for further testing. It’s currently “only answering questions,” according to Chappell.