In recent months, officials from multiple U.S. federal agencies have expressed concerns about the safety and reliability of artificial intelligence tools from Elon Musk's xAI company, according to people familiar with the matter, highlighting ongoing disagreements within the U.S. government over which AI models to deploy.
Those warnings follow the Pentagon's decision this week to allow xAI's chatbot Grok to be used in classified scenarios, putting it at the heart of some of the United States' most sensitive and top-secret operations.
A January 15 summary of the U.S. General Services Administration (GSA) report stated that Grok-4 "does not meet the security and alignment expectations required for federal government general and federal experimental AI platforms."
The full 33-page report discusses public safety incidents and the results of GSA's own testing, concluding that even limited government use of Grok requires strict, multi-level security supervision, otherwise its access "will bring higher and difficult to control security risks."
A spokesman for the General Services Administration said its assessment applies only to its own agency, and each agency will use different criteria based on its "specific business mission and risk tolerance."
Related articles:
xAI reaches agreement with Pentagon to give Grok access to classified US military systems