Pentagon Approves Elon Musk’s Grok Chatbot for Classified Use Despite Safety Alarms from Federal Agencies

The Pentagon authorizes xAI’s Grok for sensitive operations, favoring its loose controls over "woke" rivals despite warnings from federal agencies.

By: AXL Media

Published: Feb 28, 2026, 3:29 AM EST

Source: The information in this article was sourced from WSJ

Internal Government Disagreements Over AI Safety

The decision to place xAI at the center of the nation's most secretive operations follows months of internal debate among federal officials. Sources familiar with the matter indicate that multiple agencies issued warnings about the safety and reliability of Elon Musk’s artificial intelligence tools. These concerns were notably voiced prior to the Pentagon's official approval this week, highlighting a growing rift within the U.S. government regarding the appropriate standards for AI deployment in national security.

The "Woke" Debate and Political Polarization of AI

The choice between AI providers has become increasingly political, with the debate often centering on the perceived ideological leanings of different models. Senior U.S. officials, including some at the White House, have reportedly characterized rivals like Anthropic as being potentially too "woke" to be reliable. This criticism stems from Anthropic's outspoken stances on AI safety and its ties to major Democratic donors. In contrast, the Pentagon found Grok more attractive specifically because of its looser controls and Musk’s public commitment to unfiltered speech, viewing these attributes as a better fit for military requirements.

Deadlines and Competition in the Classified AI Market

Before the deal with xAI, Anthropic was the only developer approved for classified use by the U.S. military. In a move that signals a shift in the Pentagon’s leverage, the agency has reportedly given Anthropic a Friday deadline to agree to looser rules regarding its use by the military. This ultimatum suggests the Pentagon is actively seeking to lower the guardrails on its AI tools to facilitate more expansive and flexible applications in the field.