The Technology section is published with the support of Favbet Tech
Microsoft has created a generative artificial intelligence model based on GPT-4, developed specifically for US intelligence agencies. It works without an Internet connection.
This is reportedly the first time Microsoft has deployed a core language model in a secure environment, designed to allow intelligence agencies to analyze top-secret information without connection risks, as well as enable secure conversations with a chatbot similar to ChatGPT and Microsoft Copilot.
According to Bloomberg, the new artificial intelligence service (which does not yet have a public name) corresponds to growing interest among intelligence agencies in using generative AI to process classified data while reducing the risks of data leaks or hacking attempts. ChatGPT usually runs on cloud servers, which may entail risks of data leakage and interception. So last year the CIA announced its plan to create a service similar to ChatGPT, but Microsoft's solution is a separate project.
William Chappell, Microsoft's chief technology officer for strategic missions and technology, noted that development of the new system included 18 months of work modifying an AI supercomputer in Iowa. A modified GPT-4 model is designed to read files provided by users, but does not have access to the open Internet.
“This is the first time we've had a sandboxed version – where sandboxed means it's not connected to the Internet – and it's on a special network that only the government can access USA,” Chappell said.
The new service was activated on Thursday and is now available to approximately 10,000 people in the intelligence community. It is ready for further testing by relevant agencies. According to Chappell, he is now “answering questions.”
Online course “Product Design” from Skvot. Learn to create functional, trendy and ergonomic designs for furniture and interior items. More details about the course program and lecturers
One of the major disadvantages of using GPT-4 to analyze critical data is that it can potentially produce inaccurate results, draw inaccurate conclusions, or provide inaccurate information to its users. Because trained AI neural networks are not databases and operate on statistical probabilities, they can produce poor factual resources unless supplemented by external access to information from another source through a technique such as advanced search generation. Given this limitation, it is possible that GPT-4 could potentially misinform or mislead US intelligence agencies if not used properly.
Favbet Tech is IT a company with 100% Ukrainian DNA, which creates perfect services for iGaming and Betting using advanced technologies and provides access to them. Favbet Tech develops innovative software through a complex multi-component platform that can withstand enormous loads and create a unique experience for players. The IT company is part of the FAVBET group of companies.