Princeton University researchers Arvind Narayanan and Sayash Kapoor publish AI Snake Oil, dissecting mismatch between generative and predictive AI claims through clear criteria and real‐world examples to curb societal hype.

Key points

• Princeton researchers introduce AI Snake Oil with common‐sense tests to evaluate AI claims and curb hype.
• Distinguishes generative AI’s probabilistic content creation from predictive AI’s unreliable real‐world forecasting.
• Critiques futurist narratives—Kurzweil’s Singularity and Harari’s analogies—for misrepresenting AI capabilities and risks.

Why it matters: Clarifying AI hype directs attention to real technical challenges and societal impacts rather than speculative fears or exaggerated promises.