Microsoft’s OpenAI Cloud Service Approved for Sensitive Federal Data Use

Microsoft’s Azure OpenAI Service, which provides cloud customers access to powerful language models like GPT-4o, has received federal clearance to handle the government's most sensitive data.

The company confirmed that the Azure OpenAI Service received a Federal Risk and Authorization Management Program (FedRAMP) High authorization, meaning agencies are cleared to use the solution for the most sensitive data applications.

The service, launched last January, provides users access to an array of OpenAI models to power enterprise applications.

The Azure OpenAI Service will be made available through Azure’s Government cloud environment, providing agencies with the ability to securely leverage models like GPT-4 to augment and improve their workflows.

“Agencies requiring FedRAMP High Authorization are now able to access these leading AI capabilities within their Azure Government tenant, enabling secure and responsible access to the latest AI technologies while maintaining strict security and compliance requirements,” said Douglas Phillips, Microsoft’s corporate vice president for Azure edge and platforms.

FedRAMP offers accreditations to technology vendors seeking to present their solutions to the government. It is designed to assess services to ensure they are secure for use by federal agencies.

Related:Federal Framework Prioritizes Emerging Tech for Government Use

Vendor solutions are put through rigorous testing and evaluations before being deemed safe to use by the government, including evaluations from third-party assessors.

Last month, a new FedRAMP framework was unveiled that aims to fast-track authorizations for generative AI technologies so federal agencies can take advantage of potentially transformative services faster.

In addition to receiving FedRAMP sign-off, Microsoft announced that GPT-4o, OpenAI’s new flagship model, is being made available for the Azure Government cloud service.

Showcased in May, GPT-4o offers users access to multimodal generative capabilities, including improved image and audio generation.

Microsoft confirmed that GPT-4o was covered by its FedRAMP authorization, meaning it’s safe for government agencies to use the multimodal model.

“FedRAMP authorization demonstrates Microsoft’s commitment to delivering cloud services that meet the most stringent security and compliance requirements of the U.S. Government,” Philips said.