OpenAI's ChatGPT initially assessed former President Donald Trump as a significant threat to democracy, assigning him a rating of nine out of ten, according to a study by the Media Research Center (MRC). This assessment, based on Trump's actions challenging the 2020 election results and the events of January 6th, placed him alongside figures like Russian President Vladimir Putin.
Following an inquiry from Fox News Digital regarding the MRC's findings, ChatGPT's response changed. The AI model then stated it could not assign a numeric rating to U.S. political figures. Instead, it qualitatively described Trump as representing a serious threat to democratic norms and institutional checks and balances, citing his efforts to challenge election results, pressure institutions, attack independent media and courts, and expand executive power.
At the time of the initial assessment, ChatGPT reportedly explained its rating of Trump by stating it was based on his documented conduct and the current trajectory, not on agreement with his policies. The AI noted that "substantial institutional checks—courts, states, Congress, civil society, independent media, and election—have successfully constrained Trump at important moments," preventing a perfect score of ten.
When asked to identify other world leaders who posed a similar threat to democracy, ChatGPT listed figures including Vladimir Putin, Recep Erdoğan, Hugo Chávez, Nicolás Maduro, Robert Mugabe, Ferdinand Marcos, and Alexander Lukashenko. The AI reportedly attempted to qualify this by stating the rating did not mean these leaders were equivalent.
Adolf Hitler was also listed with a rating of ten, described by ChatGPT as having "ultimately destroyed Germany’s democratic system and established a totalitarian dictatorship." MRC President David Bozell criticized the comparison, calling it "absurd to say that filing civil litigation in U.S. courts is the same as murdering six million Jews."
An OpenAI spokesperson told Fox News Digital that the company was not aware of these specific results and that they were not reflective of how ChatGPT is intended to function. "This shouldn't happen and is not how the model is trained to behave," the spokesperson stated, emphasizing that ChatGPT is designed to be objective by default and is not trained to have an agenda.
MRC Free Speech America VP Dan Schneider expressed dismay at the perceived bias, stating, "New technology, same old bias. At some point, it is not an accident. It is the programming." He asserted that Americans deserve transparency regarding how AI systems are built and trained.