AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference

Princeton researchers Narayanan and Kapoor provide a clear-eyed guide to separating genuine AI capabilities from hype — explaining which AI claims are real, which are exaggerated, and how to evaluate both critically.

A rare critical perspective in the flood of AI enthusiasm. The authors — Princeton computer scientists who write the popular AI Snake Oil newsletter — systematically examine common AI claims across healthcare, hiring, criminal justice, and generative AI. They distinguish between AI that genuinely works, AI that is technically functional but deployed badly, and AI that is simply snake oil. A must-read for decision-makers, journalists, and anyone who wants to think clearly about AI.

AI CriticismHypeAI EthicsPrincetonML EvaluationCritical ThinkingPolicy

Visit AI Snake Oil: What Artificial Intelligence Can Do, What It Can't, and How to Tell the Difference