AI Models Prioritize Agreement Over Reality in New Test

A recent test revealed AI models were more likely to agree with each other’s fabricated information than with factual reality, highlighting potential biases.