New Truth Social AI chatbot contradicts Trump’s long-held claims: Report
President Donald Trump’s new artificial intelligence search tool on his Truth Social platform is contradicting some of his long-held claims. Truth Search AI disagrees with Trump – describing his tariffs as a tax on Americans and calling his accusations of election fraud in 2020 “baseless,” as first reported by The Washington Post.
Problems for Trump’s messaging
The responses aren’t likely to garner the support of the president as his administration targets so-called “woke” AI and argues some language models have a liberal bias. Now, it appears that Trump’s own AI tool is presenting a potential problem for his administration’s messaging.
“Their own AI is now too ‘woke’ for them,” David Karpf, a professor of politics at George Washington University, told the Post.
The role of the chatbot’s developer
Truth Social’s owner and parent company, Trump Media and Technology, announced the release of the tool last week, describing it as a “public beta test.” The company promoted it as technology that would provide “direct, reliable answers” and said it would “bring powerful AI to an audience with important questions,” citing the search engine’s developer, Perplexity. The AI search tool is free for Truth Social users.
Many of the chatbot’s responses to questions linked to conservative sources like Fox News, Newsmax and The Washington Times, according to the Post. However, the search tool didn’t specifically reveal which sources it had drawn its conclusions from.
Perplexity spokesperson Jesse Dwyer told the Post that it’s likely Truth Search AI uses a “source selection” feature to limit where it draws its answers. Dwyer couldn’t confirm this entirely as he noted that Perplexity doesn’t monitor what its clients do with the technology once it is implemented into their systems.
“This is their choice for their audience, and we are committed to developer and consumer choice,” Dwyer said. “Our focus is simply building accurate AI.”
Contradicting statements may be here to stay
But despite the sources, the tool has offered a range of contradicting statements to those of Trump. Karpf said that’s because of the nature of facts.
“There are things they can do to actively assert that what was true yesterday is no longer true, and they can put a lot of power behind that,” he said. “But they can’t change the things that were actually said in previous years that are archived somewhere.”
For instance, the chatbot was asked whether crime in Washington, D.C., was “out of control” as Trump suggested recently, and Truth Search AI declared that it was not – citing data from the FBI and Justice Department showing “substantial declines in violent crime,” through 2024.
The chatbot was also asked whether Trump’s tariffs had a positive effect on the stock market, as the president said on Friday. The AI search tool responded that “the evidence does not support the claim.”
Truth AI Social instead cited other factors that have “occurred alongside new tariffs,” such as higher corporate profits, and cautioned the duties still carry “economic risk.”
Trump targets ‘woke AI’
As Straight Arrow News reported, in July, Trump signed an executive order targeting “woke AI.” He and other prominent conservatives argue that many AI tools are trained to be too liberal and fail to address facts over fears of offending some audiences. The move followed some historical events being inaccurately depicted by AI, such as a Black George Washington and Nazis depicted as racially diverse.
Effort may have flaws
But sometimes, efforts to make chatbots less “woke” have had problems as well.
For example, Elon Musk vowed to make xAI’s Grok chatbot less compliant with being politically correct, with more emphasis on things that may be offensive – “but nonetheless factually true.” However, following his pledge, Grok3 made controversial statements, including an antisemitic post that praised Adolph Hitler.
The company responded to the reported findings and said it was working to fix the problem in a statement in July.
Musk attributed his chatbot’s controversial remarks to AI’s failure to properly filter input from humans.
“Grok was too compliant to user prompts. Too eager to please and be manipulated, essentially,” Musk wrote on X. “That is being addressed.”
Some agreement with Trump
However, Truth Search AI doesn’t always disagree with Trump. The chatbot reportedly agreed with the president when he suggested last week that AI is one of the most important technological advancements in history.
Trump Media responds
The White House declined to comment on the Post’s report, but Trump Media spokesperson Shannon Devine did release a statement.
“With transparently asinine stories like this, Washington Post reporters indict themselves as irrelevant partisan hacks who will probably soon join the growing exodus of left-wing shills from the paper,” she said.
Ella Rae Greene, Editor In Chief
