Microsoft has been adding AI services for personal, business, and education customers for just over a year. Today, the company confirmed that it has launched another generative AI service for a branch of the US government for the first time.
Bloomberg reports that this new AI service, which at the moment does not have an actual public name, was made for use by US intelligence organizations like the CIA. It actually launched last week but was just officially announced today.
Bloomberg quoted William Chappell, Microsoft’s chief technology officer for strategic missions and technology, as saying:
This is the first time we’ve ever had an isolated version — when isolated means it’s not connected to the internet — and it’s on a special network that’s only accessible by the U.S. government.
Chappell added that this new AI service, which is based on OpenAI"s GPT-4 LLM, can read data but cannot learn from it, in addition from the service being cut off from data on the internet. That is supposed to keep any classified data from being leaked.
While the new spy AI service from Microsoft is live right now, it will still need to be tested and approved by US intelligence groups. There"s no word yet on how long that period of testing will take. There"s also no word on what price Microsoft plans to charge the government for this spy AI service.
In March, the US House of Representatives" Office of Cybersecurity banned the use of Microsoft"s Copilot generative AI assistant by its staff members. The US House said there was a "risk to users due to the threat of leaking House data to non-House approved cloud services."
Microsoft announced plans back in November 2023 to launch a version of Copilot for government work sometime in the summer of 2024. Since then, the company has only said it still plans to launch that government-based Copilot service sometime later in 2024.