The idea behind the White House’s new artificial intelligence chatbot, called America.gov, is not inherently absurd. The apparent goal is to create an AI-driven tool to help the public with federal services, rather than directing people to try to navigate multiple government websites.
“Whether you’re a scout leader who wants to reserve a campsite at a national park, a senior citizen trying to choose a Medicare plan, or an American veteran trying to get the benefits that you’ve earned and so justly deserve, America.gov will be there to help and to help you accurately and quickly,” Donald Trump said at its official unveiling.
What’s wrong with that? In theory, nothing. In practice, though, some pitfalls quickly came into view.
For one thing, the president and his team boasted that the AI service was live, though the administration soon after conceded that it wouldn’t be fully functional until next year. For another, Trump, during his relatively brief remarks, conceded that the new service might get hacked.
But perhaps most importantly, the White House’s new AI chatbot, which the president described as “damn good,” started presenting the public with facts that were wholly at odds with Trump’s alternate reality. The New York Times reported:
According to America.gov: It is Americans, and not foreign companies, that pay for Mr. Trump’s import tariffs; U.S. elections are generally safe and secure and prosecuted cases of voter fraud are “tiny” compared to the hundreds of millions of votes cast; Mr. Trump is wrong when he says that 25 million immigrants illegally crossed the border under his predecessor, President Joseph R. Biden Jr.; Consumer prices are not going down, as Mr. Trump frequently claims.
Largely citing figures from the Congressional Budget Office, the nonpartisan scorekeeper, America.gov said that Mr. Trump’s signature domestic policy bill, signed into law last year, is estimated to add trillions of dollars to the national debt and cause more than 10 million people to lose their health insurance. The chatbot agreed with a statement that the law “on average made the poorest people in America even poorer and the richest people in America even richer.”
Journalist Helen Kennedy kept going down the same path, highlighting examples in which the new AI chatbot said that the president couldn’t use taxpayer money to pay for campaign ads; Trump was included in the Epstein files; the president lacked the unilateral authority to tear down the East Wing of the White House; Trump did not have the legal authority to rename the Kennedy Center; and the president couldn’t legally build a “triumphal” arch without congressional approval.
Oops.
In other words, the president said America.gov would help people “accurately and quickly,” seemingly unaware that he was directing the public to an official resource that presented facts he prefers to pretend don’t exist.
This was, however, the first day for the project, and it’s probably best to see it as a work in progress. Indeed, The Associated Press reported, “As Trump was still speaking at Tuesday’s launch event … the site began modifying its answers. Asked again about the 2020 election, it replied that it couldn’t answer ‘political questions.’”
Given the circumstances, it seems awfully likely that America.gov will soon be “fixed” to provide answers more in line with the president’s agenda, reality be damned.
The post On its first day, the White House’s new AI chatbot causes embarrassing troubles for Trump appeared first on MS NOW.
