Microsoft plans to open the GPT-4 large model to the U.S. Department of Defense, NASA and other agencies

June 8 news, on Wednesday, local time in the United States, Microsoft announced that it will open OpenAI’s artificial intelligence model to its Azure Government cloud computing service customers. These customers include multiple U.S. government agencies, which will have access to large language models such as GPT-3 and GPT-4.

Microsoft is OpenAI’s largest investor and uses the latter’s technology to power its Bing chatbot. Microsoft plans to announce on Wednesday that Azure Government customers will now have access to OpenAI's two big language models: GPT-4, the company's latest and most powerful model, and GPT-3, an earlier model available through Microsoft's Azure OpenAI service.

Microsoft plans to publish a blog post about the project on Wednesday. Although some media have seen the content in advance, it has not disclosed the specific names of U.S. institutions that can use the large language model by then. It is reported that the U.S. Department of Defense, the U.S. Department of Energy, and NASA are all government customers of Azure Government services.

Officials from the Defense Technical Information Center (DTIC), a subsidiary of the U.S. Department of Defense, confirmed that they will experiment with OpenAI models through new Microsoft products. This information center focuses on collecting and sharing military research results.

Microsoft already makes OpenAI models available to its commercial customers, and the Azure OpenAI service has grown rapidly in recent months. Microsoft said in May that the number of customers for the service jumped to 4,500 from 2,500 in the previous quarter, including companies such as Volvo, IKEA, Mercedes-Benz and Shell.

The initiative announced by Microsoft on Wednesday is the first time a major company has opened up chatbot technology to U.S. government agencies.

Since OpenAI publicly released the chatbot ChatGPT in late 2022, interest in large language models has soared. Large language models are trained on large amounts of Internet data and can generate human-like responses to user prompts. Since then, tech companies large and small have begun making powerful chatbots available to users, and debate has erupted in the U.S. Congress over whether and how artificial intelligence should be regulated.

Microsoft’s Bill Chappell wrote in a blog post that U.S. federal, state and local government customers can access OpenAI’s GPT-4 and GPT-3 models to perform tasks such as generative research Tasks such as answering questions, generating computer code, and compiling field reports.

A Microsoft spokesperson confirmed that while Azure Government customers will be able to use models with a chat interface, they will not have access to ChatGPT. ChatGPT is generally available through the Azure OpenAI service. Chappell previously worked at the Defense Advanced Research Projects Agency (DARPA), a research agency affiliated with the Department of Defense.

Microsoft hosts OpenAI models in its commercial cloud computing space, which is separate from the cloud services used by Azure Government customers because the latter needs to adhere to various specific security and data compliance rules.

Chappell wrote that because Microsoft hosts the models in its Azure infrastructure, any data sent to them remains in the Azure OpenAI service. The company added that data from Azure Government customers will not be used to train AI models.

The above is the detailed content of Microsoft plans to open the GPT-4 large model to the U.S. Department of Defense, NASA and other agencies. For more information, please follow other related articles on the PHP Chinese website!