Grok-maker xAI gets USDA backing in pursuit of FedRAMP High authorization

Elon Musk-led xAI is pursuing a FedRAMP High Authorization as part of the company’s efforts to expand adoption of its tool Grok across federal agencies and their workflows. 

The company’s authorization pursuit is sponsored by the U.S. Department of Agriculture, a partnership first reported Monday by Fast Company and confirmed by FedScoop through a source familiar with the matter. Typically, the metaphorical rubber stamp indicates a high level of rigor and shows that a tool is ripe for the federal government’s most sensitive workloads.

To meet the FedRAMP High requirements, xAI will need to adhere to more than 400 security controls, a third-party assessment and documentation reviews for its Grok Enterprise for Government tool. All in all, the process could last a couple of years, according to estimates from accounting firm Schellman.  

Despite the lengthy road ahead, xAI’s pursuit of the authorization is already sending signals to federal agencies that the company is serious about courting them. For some, however, Grok’s infamous reputation could precede any security standards badge it may earn. 

“I’m not surprised this is the objective given the myriad of contracts that are in play,” said Valerie Wirtschafter, a fellow in the Brookings Institution’s foreign policy, artificial intelligence and emerging technology initiative. “But if this type of authorization does go through, it will undermine public trust in the way that the federal government is adopting AI.”

One potential hang-up for Grok, Wirtschafter said, is the chatbot’s tendency to produce biased outputs. The Trump administration has emphasized its focus on rooting out “ideological bias” in AI models as part of its AI Action Plan. 

“The idea is that the federal government shouldn’t enter into contracts with companies where their LLMs have clear political agendas,” Wirtschafter said. “Grok is one that absolutely has that.”

Other concerns have risen since the tool’s introduction. The Wall Street Journal reported earlier this year that officials across multiple federal agencies had sounded the alarm about the chatbot’s safety and reliability. Lawmakers have voiced myriad concerns about the generative AI tool as well. A coalition of more than 30 civic and public interest groups also urged the federal government to bar Grok after it generated concerning outputs.

“Grok has demonstrated persistent and escalating failures related to accuracy, neutrality, and safety — including the generation of racist, antisemitic, conspiratorial, and false content, as well as sexually exploitative outputs,” the group said in February as part of its third letter to the Office of Management and Budget. “This level of corporate conduct and system failures are patently incompatible with the Administration’s own requirements for federally procured AI systems.”

Even with the criticism, some agencies have moved forward on adoption plans. 

USDA is the latest known user, backing the company’s pursuit of FedRAMP High Authorization. An Energy Department-run national laboratory is piloting the tool for document creation and general research. The Department of Health and Human Services, the Treasury Department and the Office of Personnel Management are also using the tool to varying degrees. 

Still, the chatbot is not ubiquitous. 

“When I talk to technologists who are using these tools, they’re using the tailored versions, like through StateChat, but those aren’t built on Grok,” Wirtschafter said. “ChatGPT and Claude are way more prominent.”