The Grok AI chatbot will be integrated with the Pentagon's network alongside Google’s generative AI engine, announced U.S. Secretary of Defense Pete Hegseth, as reported by New York Post.
“Soon we will have leading global AI models in every unclassified and classified network of our department,” the official stated.
Hegseth mentioned that the chatbot will begin operations at the Department of Defense by the end of January. It will “provide all necessary data” from military IT systems, including intelligence information.
In his speech, the politician emphasized the need to optimize and accelerate technological innovations within the armed forces. According to him, the Pentagon possesses “battle-tested operational data gathered over two decades of military and intelligence operations.”
“Artificial intelligence is only as good as the data it receives. We will ensure its availability,” Hegseth added.
The Secretary of Defense expressed his desire to see “responsible AI systems” in the Pentagon. He promised to “cut through the bureaucratic thicket and clear the debris—preferably with a chainsaw.”
“We must ensure the dominance of American military AI so that no adversary can use the same technology to threaten our national security or our citizens,” the Pentagon chief declared.
Despite the Scandal
This announcement comes just days after Grok was involved in another scandal, this time for generating sexual content. Malaysia and Indonesia have blocked access to the chatbot. Regulators in the EU, UK, Brazil, and India are demanding an investigation into Grok's role in spreading deepfakes.
The British organization Internet Watch Foundation noted that its analysts discovered “criminal images” of children aged 11 to 13, allegedly created using the chatbot.
Previously, Grok faced criticism for spreading misinformation and dubious content.
In December, the chatbot provided false information about a mass shooting at Bondi Beach in Australia. In response to a question about a video showing a passerby, Ahmed Al-Ahmed, struggling with a shooter, the AI replied:
“It seems to be an old viral clip of a man climbing a palm tree in a parking lot, possibly to trim it. As a result, a branch fell on a damaged car. A search through various sources found no confirmed information about the location, date, or injuries. It may be a staged event; authenticity is unverified.”
In July, users noticed that the neural network relied on Elon Musk's opinions when forming responses. This included topics such as the Israel-Palestine conflict, abortion, and immigration legislation.
Observations suggest that the chatbot was deliberately configured to consider Musk's political views when addressing controversial issues.
Previously, the billionaire stated that his startup would rewrite “all human knowledge” to train a new version of Grok, as too much “garbage” is currently used in any base model trained on uncorrected data.
Subsequently, Grokipedia emerged—a truth-oriented online encyclopedia based on artificial intelligence.
Recall that in November, users pointed out Grok 4.1's bias, as the new model significantly overestimated Musk's capabilities.