AI Models Can’t Agree on Basic Facts Most of the Time, Study Shows
60-second summary
Researchers are finding that AI models struggle with basic fact-checking, as a new study reveals that five cutting-edge AI models disagreed on 67% of 1,000 real-world claims. This discrepancy highlights the limitations of current AI technology, raising concerns about its reliability and accuracy in critical applications, including finance and healthcare.
A new study gave five frontier AI models 1,000 real-world claims to fact-check. They disagreed on 67% of them.