Pentagon embraces Elon Musk’s Grok AI despite internal warnings over safety risks
The Pentagon approved Elon Musk’s Grok AI for classified military operations, despite warnings from agencies like the GSA and NSA about its reliability, susceptibility to bias and security risks.
Internal memos raised alarms over Grok’s vulnerabilities, including sycophantic responses and susceptibility to manipulated data. Despite assurances from xAI executives, Grok remains excluded from the GSA’s AI testing platform (USAi) due to unresolved risks.
The Pentagon favors Grok’s free-speech, unfiltered approach, contrasting with rivals like Anthropic’s Claude, which faces criticism for being “too woke” due to strict content moderation. The NSA, however, flagged Grok as uniquely risky compared to competitors.
The Pentagon’s push for Grok coincided with the resignation of its chief of responsible AI, Matthew Johnson, who warned that safety protocols were being sidelined in favor of accelerated military AI adoption.
The Pentagon’s gamble on Grok reflects the broader U.S.-China AI rivalry, with military strategists valuing its adversarial simulation capabilities—despite critics warning that weak guardrails could compromise national security.
Read Full Article: https://www.naturalnews.com/2026-03-02-pentagon-embraces-grok-ai-despite-safety-risks-warnings.html
