Artificial intelligence’s tendency to agree with users, rather than its propensity to fabricate information, may pose a significant risk. This…