Two House Democrats are pressing the White House to hand over President Donald Trump’s Grok chat history, arguing that his conversations with artificial intelligence chatbots — including prompts about deposing Venezuela’s former president Nicolás Maduro — belong to the public. In a joint letter sent to Trump last week, Rep. Jamie Raskin, a Maryland Democrat, and Rep. Robert Garcia, a California Democrat, said the president’s AI prompts and the responses he received should be treated as public records under the 1978 Presidential Records Act, according to CNET.
The lawmakers also raised ethical and security objections to what they described as the president’s reliance on “unsecured, privately operated AI systems” to carry out foreign policy decisions, alleging that Trump consulted Grok, the chatbot built by Elon Musk’s xAI, but not Congress, on the military deployment to Venezuela.
What the Grok Chat History Reportedly Shows
According to public reporting cited in the letter, Trump spent multiple hours in December 2025 prompting Grok about how Venezuelans would respond to Maduro’s capture. That was just weeks before he ordered the military operation and Maduro’s arrest. Trump reportedly “came away thinking Grok was ingenious.”
Officials alleged that the chatbot told the president “that Maduro was a repressive and deeply unpopular dictator and that many Venezuelans would likely celebrate his downfall.” In the wake of the invasion, CNET reported, X users continued to prompt Grok about the operation, asking whether the South American country “deserved” to be bombed. The country’s post-operation economy has since drawn attention of its own, including 100-year oil field concessions granted to Blue Energy.
Raskin and Garcia Invoke the Presidential Records Act
Garcia and Raskin argue in the letter that Trump’s use of AI “exposes the country to malign influence from unknown and unaccountable third parties.” They have asked the White House to explain whether it restricts classified information from AI chatbots, and whether vendors, subcontractors or other external parties can access prompts, outputs, logs or metadata.
The two members used blunt language about the chatbot’s owner, referring to Grok’s antisemitic episode in 2025 in which the system referred to itself as “MechaHitler.”
“If the American President is relying on the policy advice of a Musk-designed, Hitler-friendly chatbot with a predilection for child pornography in order to carry out his official duties, Americans have a right to know,” the joint letter reads.
The letter also addresses the White House’s ongoing use of AI chatbots in military operations abroad, including the war with Iran: “If the United States military killed dozens of innocent schoolchildren based on the faulty advice of an AI chatbot, we have a right to know that too.” In June, months after a US military strike on an Iranian elementary school killed at least 123 children, the Pentagon’s artificial intelligence chief said Grok’s continued operation was “a matter of paramount national security.”
Separately, Grok’s outputs are already the subject of litigation, including a case in which Tennessee girls sued Grok and Stability AI over deepfakes.
A Federal Report Flagged Grok as a Security Risk
A January executive summary of a General Services Administration report, reviewed by The Wall Street Journal, concluded that xAI’s Grok “does not meet the safety and alignment expectations required” for general federal use and may pose a security risk in classified settings.
CNET noted that using a consumer AI product rather than a hardened government system means those chats could be processed by a private company’s infrastructure. The outlet also pointed to the question of Musk’s political influence over a system he owns and has repeatedly altered to reduce content filters.
Sycophancy, Hallucination and a Misidentified Chinese Vessel
CNET reported that large language models are probabilistic text generators that can sound authoritative while producing biased or fabricated information, and that models are known to amplify existing biases by agreeing with users’ strongly held beliefs. One journalist found that previous Grok models had “tripled” sycophancy rates compared with popular ChatGPT and Claude models, representing a shift in “behavior that MIT researchers identify as actively harmful.” Grok models are also trained on public interactions from X.
Hallucination has already reached a military decision chain. Earlier this year, according to reporting cited in the letter, US military officials received an intelligence report alleging that a Chinese vessel in the Middle East was carrying components for a nuclear weapons program. Only as the military prepared to intercept the ship — an action that many say could have escalated into a confrontation with China — did officials discover that the report’s author had relied on an AI chatbot and misidentified the cargo.
Pentagon Contracts and Operation Epic Fury
The Pentagon said in a sworn statement that xAI’s latest models enabled US forces to deploy over 2,000 munitions to target 2,000 different locations within 96 hours during Operation Epic Fury, the US campaign against Iran. Palantir’s Maven Smart System, which integrates intelligence from satellites, drones and other sources to identify military targets, reportedly played a central role in that campaign as well. The Defense Department has also awarded contracts worth up to $200 million each to Anthropic, Google, OpenAI and xAI to accelerate military adoption of advanced AI.
CNET noted that researchers have regularly pointed to Israel’s bombing of Gaza as an example of AI-first warfighting. According to analysis cited by the Georgetown Security Studies Review, Israel’s Lavender system marked roughly 37,000 Palestinians as suspected Hamas or Islamic Jihad operatives early in the war, with an estimated error rate of 10%. The system “dramatically expanded” Israel’s kill list by relying on signals such as communication patterns and group-chat membership, producing misidentifications that included Palestinian police, aid workers, relatives of militants and people with similar names or nicknames. A related system, “Where’s Daddy?”, tracked marked Palestinians and alerted the military when they entered their homes, striking multiple private family residences.
The Nov. 11 Briefing Deadline
Raskin and Garcia are requesting a full description of the president’s personal use of the technology and a closed-door briefing by Nov. 11 to review the administration’s AI protocols and cybersecurity safeguards. Treating the Grok chat history as official government property under the Presidential Records Act would make those AI conversations public records, which CNET reported would open the door for Congress to subpoena the executive branch.
Reached for comment, the White House referred CNET to the Pentagon, which didn’t immediately respond. Rep. Garcia’s office didn’t respond to a request for comment.
