Deepfake Fears Explode After Trump’s AI Chat

When a sitting president chats with an AI version of Theodore Roosevelt, it highlights how technology and politics are merging in ways that worry Americans who already feel the system is slipping out of their control.

Story Snapshot

President Trump visited the new Theodore Roosevelt Presidential Library and spoke with an AI hologram of Roosevelt.
The Roosevelt avatar is powered by a private AI system trained on verified historical sources, not the open internet.
Trump later retold the encounter in public, raising questions about whether he changed some details of the exchange.
This high-tech “conversation with history” lands in a time when many Americans distrust both government and artificial intelligence.

Trump’s visit and his “conversation” with Roosevelt

On July 1, 2026, President Donald Trump traveled to Medora, North Dakota, to help dedicate the new Theodore Roosevelt Presidential Library. During the visit, Trump toured exhibits that recreate White House spaces and then interacted with an artificial intelligence hologram of former President Theodore Roosevelt. Video clips across Facebook, Instagram, and YouTube show Trump facing a life-like Roosevelt avatar, speaking and asking questions as if the two presidents were meeting across time. The moment was presented as both historic and futuristic.

The hologram did more than play back canned lines. The library used artificial intelligence to create a Roosevelt avatar that responds to visitors in real time, including the president. In one clip, Trump asks Roosevelt whether he considers the Panama Canal his greatest achievement, and the avatar answers from a scripted historical point of view. The exchange fits a larger America250 push by federal agencies to use digital tools to make history feel immediate to visitors and to tie modern leaders to past icons. Supporters call it engaging; skeptics see political theater with a silicon mask.

How the AI Roosevelt actually works

The Roosevelt hologram is not a random chatbot talking off the top of its digital head. Technology partner Terentia describes the system as a private “GPT engine” trained on a large language model built only from Roosevelt-related materials. Those materials include primary sources from 33 collections and 18 institutions, plus carefully chosen biographies to balance bias. The goal is for the avatar to act as a kind of virtual historian, linking answers back to diaries, letters, and speeches, rather than inventing facts or repeating internet gossip. Sensitive personal family letters are blocked from the system to protect privacy and dignity.

Library leaders had previewed this idea for months. Executive Director Robbie Lauf said visitors would “engage and talk with Theodore Roosevelt through the power of artificial intelligence” about his life and about modern issues facing the country. The Interior Department has promoted a similar Roosevelt avatar for Theodore Roosevelt National Park as part of the America250 celebration, promising a human-like figure that answers visitor questions using advanced artificial intelligence tools. Together, these projects show how elites in government, tech, and culture are reshaping how citizens meet the past—through filters they design and control.

Trump’s retelling and the question of accuracy

After the visit, Trump publicly described his exchange with the AI Roosevelt and highlighted dramatic details from Roosevelt’s life, including a diary entry from 1884 describing the deaths of both his wife and his mother on the same day in the same house. That diary event is real history, but available sources do not yet include a full transcript of what the avatar actually said to Trump in the room. A widely shared Facebook post claimed Trump “appeared to alter details” of his exchange when he later recounted it, though it did not provide full audio of the original conversation. That claim feeds a familiar pattern where critics question Trump’s honesty, while supporters distrust the media’s framing.

The lack of an official transcript or full unedited video leaves a gap between what happened and what voters are asked to believe. Short clips show Trump’s side of the dialogue, but they may be edited or cut for effect. That fragmentation matters in a country where many citizens already think the “deep state” and political media class spin events to fit their own stories. When technology mediates the past and then partisan outlets mediate the technology, people on both the right and the left can feel one more layer of control slipping away from them.

AI history, political messaging, and public trust

Trump’s chat with AI Roosevelt is not just a museum stunt; it sits inside a wider shift where artificial intelligence shapes how people learn history and how they feel about politics. Research shows that factually accurate AI summaries of historical events can still nudge readers’ political views depending on how the facts are framed. In one study, default chatbot summaries pushed opinions slightly more liberal than neutral Wikipedia articles, while tuned conservative summaries moved some readers right. That means the way a digital Roosevelt talks about big business, war, or conservation could quietly tilt attitudes.

He thinks the AI generated video he watched of Roosevelt was actually the real "Theodore Roosevelt". https://t.co/DdMkzXq4kh

— Levin Inches™ (@LevinInches) July 2, 2026

Global reports warn that generative artificial intelligence is already used to manipulate voters, create fake clips of leaders, and manufacture a sense of consensus around contested issues. At the same time, real clips of Trump have been rebranded as “AI fakes” to discredit them. So Americans who already feel betrayed by elites now face a new problem: it is harder to know when you are hearing a real voice, a reconstructed one, or someone’s spin on both. The Roosevelt avatar is built for education, but any tool that makes history speak on command can be used to bless today’s policies as if they carry the seal of the past.