I’m a Professional Fact-Checker. AI Is Wrong More Often Than You Think

A WIRED fact-checker says human verification is still essential because AI search tools and chatbots remain frequently inaccurate.

The article cites studies from Full Fact, the Tow Center, the BBC, and arXiv showing error rates that can range from about one-third to nearly half or more.

AI Overviews and other LLM-based systems can produce confident but wrong answers, especially on complex or current topics.

The takeaway: journalism, research, and misinformation work still need trained human fact-checkers to confirm AI-generated claims.