skeptical prompting

Skeptical prompting is a technique for interacting with AI that encourages the model to disagree, show its reasoning, and surface hidden assumptions. Rather than treating large language models as agreeable assistants, skeptical prompting pushes them to challenge user inputs and expose potential flaws. This approach is especially relevant for professionals using generative AI in drafting, analysis, or automation, where conversational fluency can mask errors or biases. The concept, introduced by psychology professor turned AI consultant Leigh Coney, aims to improve decision-making by forcing AI to reveal its logic and uncertainties before users act on its outputs.