Although researchers claim that AI chatbots are real-time suppliers and arbiters of news, the past nine days in American politics have proven otherwise.
According to the Washington Post Reports, from the assassination attempt on Donald Trump to President Joe Biden abruptly dropping out of the presidential race, AI chatbots were not up-to-date, refused to tell the answer, and didn’t even accurately report events.
Following the July 13 shooting at former President Donald Trump’s rally in Butler, Pennsylvania, many chatbots were confused about what had transpired. However, ChatGPT called the rumors of assassination misinformation, while Meta’s AI said it didn’t have any information about such an event.
Chatbots’ situation was similar when Trump named J.D. Vance as his running mate and when President Biden tested positive for the coronavirus on Wednesday.
The reports showed that Microsoft Copilot was comparably better at reporting accurate breaking news. However, it was also left stumped when asked who was running for U.S. president and said, “Looks like I can’t respond to this topic.”
When ChatGPT was asked about Biden’s withdrawal, even after one hour of the event, it claimed he was still running.
The reason behind this lack of answer can be due to Microsoft, Meta, and Google’s plans to restrict some election-related queries on their AI tools, and thus, their bots refuse to reply to any political query.
This research indicated that AI chatbots and their outputs can be huge liabilities. However, the point is that if AIs can’t provide accurate information in real-time, then why is it so overblown?
Jevin West, a professor and co-founder of the Center for an Informed Public and the University of Washington, said, “They’re incredibly good at communication. They’re not being optimized necessarily for truth.”
