Report shows Grok is rarely used in US federal government records

In short: A new Reuters review suggests Elon Musk’s Grok chatbot is barely being used across recorded US federal government AI work.

What's going on

Reuters reviewed more than 400 examples of US government AI use where an AI vendor was named. Grok, or Musk’s company xAI, appeared only three times. Those mentions were for basic tasks like drafting documents or managing social media, and Grok was listed alongside other tools.

By comparison, OpenAI’s models appeared in more than 230 examples. Reuters found Google and Anthropic were also named many more times than Grok. Another government database covering smaller, more experimental AI projects showed a similar pattern, with Grok again appearing only three times.

The data is not a complete picture of all government AI use. Many records do not name a vendor, and agencies may define “AI” differently. The lists also do not include intelligence agencies or the Pentagon, where xAI previously received a $200 million contract and was cleared to operate on classified networks.

What to watch

The low visibility of Grok in these records matters because Musk and SpaceX have been tying Grok closely to SpaceX’s business story ahead of a possible IPO, meaning a stock market launch. If Grok is not widely chosen for day-to-day work, it could raise questions about how strongly it will sell to other large organizations. Think of it like a new office assistant tool, if most departments keep picking other brands, it can be hard to argue it should be central to the company’s future plans.

Source: The Verge AI