People are less likely to believe an AI if it conveys uncertainty

When a large language model expresses doubt about the information it supplies, people are less likely to accept it as fact and more likely to find accurate information elsewhere