
An AI chatbot launched by the US government on America.gov contradicted Donald Trump on the 2020 election on its first day, then stopped answering political questions the next day, according to statements from the White House and the Digital Democracy Institute of the Americas.
AI-generated summary
The US government has launched America.gov, an AI chatbot intended to facilitate access to federal information and services, presented as a tool to combat disinformation.
A US government chatbot powered by artificial intelligence (AI) has stopped answering some political questions, just one day after it went live. “Hello, America. Whatever your needs from the government, start here,” we can read on the America.gov platform, which aims to simplify access to government information and federal services.
The American president described the tool as “damn good”. But the chatbot quickly contradicted Donald Trump on certain subjects. So, when asked on Tuesday whether there was widespread voter fraud in the 2020 election, which the current president claims to have won, the chatbot responded: “Official federal assessments conducted after the 2020 presidential election did not find fraud on a scale that would have changed the outcome.”
Skip the ad
“From its first day, America.gov has relied on facts and official documents to contradict some of the false narratives and political interpretations regarding immigration and elections widely relayed by the administration that created it,” notes the Digital Democracy Institute of the Americas. “Early analyzes showed that America.gov generally provided evidence-based responses supported by government documents, even when those documents contradicted claims by Trump and his administration,” the institute continues. The chatbot's response regarding the 2020 election was quickly taken up by elected Democrats. “Finally, a clear answer in the Trump administration,” Sen. Adam Schiff wrote Tuesday on X.
Change of answer
But the tone of the tool changed on Wednesday. “I do not provide political commentary or analysis of past elections,” he responded to the same question. However, he continued to refute other contested assertions by Donald Trump, on inflation for example.
It is difficult to know at this stage whether the government has changed the chatbot's settings. “America.gov will continue to be updated to provide Americans with the most accurate and reliable federal information and resources,” the White House said in a statement to AFP.
AI outlook — possibilities, not facts
The US government will issue clear guidance on the expected political neutrality of AI-based communications tools.
Likely · Within weeks

A research paper by many of Silicon Valley's top engineers warns of recursive self-improvement in artificial intelligence, a concept that allows AIs to create their own successors at an uncontrollable pace, dubbed an "intelligence explosion."

Professionals like graphic designer Babeth Reynaud and consultant Aurélie Bécat are losing clients because of AI and are retraining for less exposed manual professions, while a study estimates that five million jobs could be threatened in France.

According to a NewsGuard report, fake articles and reports published by pro-Kremlin actors imitating recognized media outlets have exploded by 850% over one year. Because of its support for Ukraine, France is a favored target.

Facing the giants of Silicon Valley, Swiss hosting company Infomaniak is developing Euria, an AI assistant focused on data confidentiality, local hosting and ecological home heating.

In a shocking research paper, AI engineers including OpenAI's Jakub Pachocki, Geoffrey Hinton and Yoshua Bengio warn that self-recursive improvement (RSI) could lead to an "intelligence explosion", where AI improves on its own without human help, becoming a strategic priority for the sector.

Donald Trump said Tuesday that leaders of big tech companies had reached a moral agreement to self-regulate artificial intelligence, during a meeting at the White House with figures including OpenAI's Greg Brockman, Anthropic's Dario Amodei, Elon Musk and Jensen Huang. He also announced that he would sign a document establishing the term "super intelligence" in place of "artificial intelligence."