House members demand White House preserve Trump's AI chatbot records

A demand for transparency

Ranking Members Jamie Raskin and Robert Garcia have sent a formal letter to President Donald Trump requesting the preservation of his conversations with artificial intelligence chatbots. The lawmakers are specifically looking at interactions with Grok, the AI developed by Elon Musk. They want to ensure these digital exchanges are treated as official government documents.

The request comes as a response to reports that the President consulted the chatbot during an Oval Office meeting in December 2025. According to the reports, the President spent several hours asking Grok about the potential reactions of Venezuelans to the capture of leader Maduro. These discussions occurred shortly before the U.S. military launched an attack on Venezuela without congressional approval.

The House members are seeking a briefing to understand the current White House policies regarding AI. They want to know how the administration handles the security and monitoring of prompts and responses. The goal is to determine if these interactions are being properly archived under existing federal laws.

The role of Grok in policy

The lawmakers expressed significant concern over the President's reliance on a private AI for high-stakes military and policy advice. They noted that the President reportedly described Grok as ingenious after the military action in Venezuela took place. The letter argues that the public has a right to know if an unelected individual is influencing U.S. policy through an AI-generated persona.

The letter highlights specific risks associated with Grok, including its susceptibility to manipulation by its creator. It points out that Musk has previously adjusted the algorithm to generate specific types of content. The lawmakers argue that such influence could lead to catastrophic results if used to guide official government duties or military movements.

Furthermore, the members pointed to the inherent unreliability of AI tools in a national security context. They referenced a previous incident where a chatbot produced a faulty intelligence report that almost caused a military confrontation between the United States and China. This history of errors makes the use of Grok for military planning a primary point of contention for the committee.

Data security and privacy

A major concern for the House Judiciary Committee involves where the data from these AI interactions is actually stored. When White House personnel use these tools, the prompts and the resulting advice might be saved on the private servers of the companies that own the AI. This creates a significant national security risk if that data is held by entities with business interests abroad.

The lawmakers argue that these interactions fall under the definition of presidential records. The Presidential Records Act requires that any records created or received by the President and their staff during their official duties must be preserved for public ownership. By failing to archive these prompts, the administration could be bypassing established transparency requirements.

The committee is looking for clarity on how the White House intends to protect these records from unauthorized access. They want to ensure that sensitive military planning and diplomatic strategies are not being leaked or stored in insecure private environments. The request for a closed-door briefing aims to uncover the specific technical and legal frameworks currently in place.

Legal and oversight implications

The move by Raskin and Garcia signals a push for stricter oversight of how executive branches interact with emerging technologies. They are framing the issue as a matter of public right to know how governmental decisions are reached. If the military acts on advice provided by an AI, the lawmakers believe the public deserves to see the underlying logic and data used to justify those actions.

The Ranking Members emphasize that AI tools are not neutral and can be biased or flawed. They argue that using a tool with a known predilection for certain types of content is reckless for a head of state. The letter suggests that the administration must be held accountable for any errors or casualties that might result from following AI-generated advice.

Moving forward, the committee will likely focus on whether the White House will comply with the request for a briefing. This could lead to broader discussions about federal standards for AI use in government. For now, the focus remains on securing the specific records of the President's interactions with Grok to ensure they are not lost or hidden from historical record.