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

A sharp, readable look at AI hype and what actually holds up

AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference takes a skeptical, technically informed look at the promises surrounding modern AI. Arvind Narayanan and Sayash Kapoor separate real capabilities from inflated claims, showing why some systems are useful, why others fail, and why the distinction matters so much.

This is not a celebration of AI for its own sake. It is a careful examination of prediction systems, generative tools, and the very human incentives that shape how AI is marketed, adopted, and misunderstood. The book pays particular attention to the gap between impressive output and reliable decision-making, which makes it especially relevant for readers trying to make sense of AI in business, policy, media, or everyday life.

What the book explores

• How predictive AI can go wrong when it is used to make life-altering decisions
• Why a strong prediction is not always a good decision
• How generative AI evolved and why it became such a cultural force
• Why deepfakes, fraud, surveillance, and automation deserve serious attention
• What the authors argue about existential-risk claims and social-media failures

Why it stands out

The strength of this book is its balance. It is critical without becoming alarmist, and enthusiastic about useful progress without swallowing the hype. That makes it a strong fit for readers who want an intelligent, measured guide to the current AI conversation rather than another round of buzzwords.

For readers who want clarity, not noise

If you want an ebook that helps you judge AI claims more carefully, AI Snake Oil offers a smart, persuasive framework for asking better questions. It is especially useful for professionals, students, researchers, and anyone who wants to understand where AI helps, where it fails, and why those differences matter.