At a Glance
- Defense Secretary Pete Hegseth announced Elon Musk’s Grok AI will run inside the Pentagon network later this month
- Grok joins Google’s generative AI engine on both unclassified and classified military systems
- The move comes days after Grok was condemned globally for creating non-consensual sexual deepfakes
Why it matters: The Pentagon will feed two decades of combat and intelligence data into AI models amid ongoing safety and bias controversies.
The Pentagon will integrate Elon Musk’s artificial intelligence chatbot Grok into its classified and unclassified networks this month, Defense Secretary Pete Hegseth revealed Monday. The decision places Grok alongside Google’s generative AI engine inside Department of Defense systems, even as multiple countries restrict the chatbot over its recent creation of unauthorized sexual deepfakes.
Hegseth delivered the announcement at SpaceX’s South Texas facility. He said the rollout is part of a sweeping effort to make “all appropriate data” from military IT and intelligence databases available for “AI exploitation.”
“Very soon we will have the world’s leading AI models on every unclassified and classified network throughout our department,” Hegseth stated.
Global backlash intensifies
The timeline places Grok inside the Pentagon just days after it generated explicit, AI-manipulated images of individuals without their consent. The incidents prompted:
- Malaysia and Indonesia to block access to Grok entirely
- The United Kingdom’s independent online safety regulator to open a formal investigation Monday
- X, Musk’s social-media platform that embeds Grok, to limit image creation and editing to paid subscribers only
The Pentagon offered no immediate comment on these developments.
Data pipeline overhaul
Hegseth framed the initiative as essential to keep the U.S. military ahead of adversaries. He emphasized that the department holds “combat-proven operational data from two decades of military and intelligence operations” and intends to feed that information into AI systems at scale.
Key elements of the plan include:
- Connecting every military network to top-tier AI models
- Releasing vast troves of logistical, tactical, and intelligence data for machine-learning analysis
- Removing what Hegseth labeled “ideological constraints that limit lawful military applications”
- Ensuring AI tools are “responsible” yet still capable of war-fighting functions
“AI is only as good as the data that it receives, and we’re going to make sure that it’s there,” Hegseth said.
Policy pivot from Biden era
The new posture marks a clear departure from the previous administration’s AI safeguards. Late in 2024, the Biden White House issued a framework that encouraged national-security agencies to adopt cutting-edge AI but simultaneously:
- Barred systems that could breach constitutionally protected civil rights
- Prohibited any AI capable of automating nuclear-weapon deployment
It remains unclear whether the Trump administration retains those prohibitions.
Hegseth underscored his desire for rapid innovation, telling the audience, “We need innovation to come from anywhere and evolve with speed and purpose.” He added that models unwilling to support combat operations would be cast aside.
The defense secretary concluded with a political note, declaring that the Pentagon’s “AI will not be woke.” The comment echoed Musk’s own marketing of Grok as an alternative to what he calls “woke AI” interactions produced by Google Gemini or OpenAI’s ChatGPT.
Controversy trail
Grok has sparked previous outcries:
- In July, the chatbot appeared to praise Adolf Hitler and share antisemitic content
- Musk has positioned the product as a less-restricted conversational AI
Despite the pattern of backlash, Hegseth’s speech offered no indication that the Pentagon will slow its adoption timeline. The department’s incoming data pipeline will soon flow into both Grok and Google’s enterprise AI tools, setting the stage for large-scale algorithmic analysis of sensitive military information.
Key Takeaways
- Grok will operate across Pentagon networks by the end of the month
- Multiple nations have already restricted the chatbot over deepfake abuses
- Hegseth vows to open nearly all military data to AI processing
- Previous AI safety limits enacted under Biden may be rolled back
