Cloud

Microsoft launches GPT-enabled Azure AI for US government agencies


Microsoft has announced the availability of a new Azure OpenAI service, designed specifically to meet stringent requirements of US government agencies.

These generative AI capabilities offered by the new service can help US government agencies improve efficiency, enhance productivity, and unlock new insights from their data, the company said in a blog post.

“Currently, large language models that power generative AI tools live in the commercial cloud. For government customers, Microsoft has developed a new architecture that enables government agencies to securely access the large language models in the commercial environment from Azure Government allowing those users to maintain the stringent security requirements necessary for government cloud operations,” it said.

Azure Government is a specialized cloud service for US government agencies with dedicated instances that can only be used by screened users.  

“Azure Government peers directly with the commercial Azure network and doesn’t peer directly with the public internet or the Microsoft corporate network,” the company said.

In order to access large language models such as GPT-4 and GPT-3.5, as well as embeddings — numerical representations of information and relationships among text strings — government bodies can use the APIs bundled with the OpenAI service.

“You can adapt these models to your specific task, including but not limited to content generation, summarization, semantic search, and natural language-to-code translation,” Microsoft said in the post, adding that developers could use the service to build generative AI-based applications faster.

The Azure OpenAI service can also be used through the Python software development kit (SDK) or through the company’s web-based interface in Azure AI Studio, the company said, adding that data from these government agencies will not be used to train its large language models.

Copyright © 2023 IDG Communications, Inc.



READ SOURCE

This website uses cookies. By continuing to use this site, you accept our use of cookies.