Microsoft launched an AI-based chatbot for US intelligence agencies

Microsoft has launched an AI-based chatbot for American intelligence agencies. This is the first time the language model has been used offline. US CIA employees plan to use this model to analyze confidential data.

Info May 11, 2024 0 99 Add to Reading List

Microsoft launched an AI-based chatbot for US intelligence agencies

Microsoft has developed a new generative AI model for the US CIA. It is completely isolated from the Internet and is capable of processing top-secret information without any risk of leakage or hacking.

While most existing AI models, such as OpenAI's ChatGPT, typically rely on cloud services to train and extract information from data, Microsoft has offered intelligence agencies its own secure system.

The new model responds to growing interest among intelligence agencies in using generative AI to process sensitive data while minimizing the risk of data compromise.

William Chappell, Microsoft's chief technology officer, said it took 18 months to develop the AI supercomputer in Iowa. He said the model is based on GPT-4 and can answer questions. In addition, the chatbot can read and analyze files, but cannot learn to prevent sensitive information from entering the platform.

Starting today, this model has become available for use by members of the American intelligence services. The chatbot has yet to be tested and accredited by intelligence agencies.