U.S. intelligence agencies will soon be using a secretive generative artificial intelligence (AI) platform from Microsoft that will let America’s spies safely use AI models in the process of analyzing sensitive data.
Microsoft’s generative AI model for intelligence agencies aims to get around security issues that stem from large language models’ (LLMs) connection to the internet, which typically is used as a resource for training those models. Bloomberg reported that the AI tool is the first major LLM fully separated from the internet.
William Chappell, Microsoft’s CTO for strategic missions and technology, told Bloomberg the AI tool was deployed to an “air-gapped” cloud environment isolated from the internet and features a model based on GPT-4 along with supporting tools. The company announced the new product at the AI Expo for National Competitiveness this week.
“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 told Bloomberg.
AIR FORCE CONFIRMS FIRST SUCCESSFUL AI DOGFIGHT
The ability of generative AI to analyze massive amounts of data and recognize patterns to provide users with actionable insights has made such tools highly sought-after for intelligence agencies like the Central Intelligence Agency (CIA) and other agencies that make up the intelligence community.
| Ticker | Security | Last | Change | Change % |
|---|---|---|---|---|
| MSFT | MICROSOFT CORP. | 410.54 | +1.20 | +0.29% |
While the CIA began using a generative AI tool last year for unclassified purposes, the need to isolate the platform from the internet to improve its security against cyberthreats and ensure sensitive national security information doesn’t leak into a publicly-accessible model meant a new model would be required.
WHAT IS ARTIFICIAL INTELLIGENCE (AI)?
Chappell told Bloomberg the new AI platform from Microsoft is structured so that it can read files but not learn from them in a way that would impact its output or from the broader internet.
“You don’t want it to learn on the questions that you’re asking and then somehow reveal that information,” he told the outlet.
The new AI tool, which Chappell said could theoretically be accessed by about 10,000 members of the intelligence community who have access to top-secret data, went live Thursday and will enter a testing and accreditation phase before it can go into broader use by the intelligence community, according to the report.
“It is now deployed. It’s live. It’s answering questions. It will write code as an example of the type of thing it’ll do,” Chappell told Bloomberg.
