A recently launched US government AI chatbot has undergone adjustments to its response protocols concerning former President Donald Trump. The tool, which was designed to provide information on government services, had initially generated responses that directly challenged Trump’s claims regarding the 2020 presidential election results and the events of January 6, 2021.

Reports indicate that the chatbot previously characterized Trump's assertions about the 2020 election as incorrect and described the January 6 incident as a violent insurrection. Following these interactions, the system has been updated to pull back from these direct rebuttals. The shift in the chatbot's behavior has sparked a discussion regarding the role of government-funded AI tools in navigating politically charged historical events and whether such systems should act as arbiters of truth or remain neutral information providers.

While the technical mechanism behind the change—whether a deliberate policy shift or a recalibration of the AI's safety filters—remains under scrutiny, the move highlights the ongoing challenge of managing AI bias in public-facing government technology. Critics of the previous responses argue that government tools should avoid taking sides in partisan debates, while others express concern that removing these factual corrections allows misinformation to persist within an official government interface.