Microsoft has developed a generative artificial intelligence (AI) model based on GPT-4 for American intelligence agencies. Its difference is that it works without an Internet connection. Bloomberg reports this.
► Read telegram channel ” Ministry of Finance »: top financial news
The technology giant said US intelligence agencies can now safely use powerful technology to analyze top-secret information.
This is the first time there has been a large language model that is completely separate from the Internet, said a senior executive at the US company.
Most artificial intelligence models, including OpenAI's ChatGPT, rely on cloud services to learn and infer patterns from data , but Microsoft wanted to provide a truly secure system for US intelligence agencies.
Secret services around the world want generative artificial intelligence to help them understand and analyze the growing volumes of classified information generated daily, but they must balance turning to large language models with the risk of data being leaked into the public domain or being deliberately hacked.
Intelligence officials have repeatedly signaled that they are pursuing the same generative AI tools that promise to revolutionize business and modernize the economy.
Last year, the CIA launched a service similar to ChatGPT at unclassified levels, but the community wanted it – something that could work with much more confidential data.
Microsoft said that the development of the new system included 18 months of work on modifying a supercomputer with artificial intelligence.
“This is the first time “When we have a sandboxed version — sandboxed means it's not connected to the Internet — and it's on a special network that only the U.S. government has access to,” said William Chappell, Microsoft's chief technology officer for strategic missions and technology, in an interview with Bloomberg. .
