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

AI Snake Oil: How to Separate Real AI from Hype, Misinformation, and Bogus Claims

Artificial intelligence is everywhere, but so are AI hype, misleading claims, exaggerated promises, and misinformation. AI Snake Oil by computer scientists Arvind Narayanan and Sayash Kapoor, both recognized among TIME’s 100 Most Influential People in AI, offers a clear and practical guide to understanding what AI can actually do—and where its limitations begin.

The book is designed for readers who want to understand how AI works, why it often fails, where it can be useful, where it can cause harm, and how to recognize products that are being sold primarily through artificial intelligence hype.

Narayanan and Kapoor use the term “AI snake oil” to describe AI products and services that make impressive claims without reliable evidence that they actually work. These systems are often marketed as intelligent, predictive, or transformative even when their real-world performance is weak, unproven, or fundamentally limited.

While acknowledging the genuine potential of technologies such as ChatGPT and generative AI, AI Snake Oil challenges many of the exaggerated claims surrounding artificial intelligence. The authors show how AI is already being misused or oversold in areas including education, healthcare, hiring, banking, insurance, and criminal justice.

A major theme of the book is that not all forms of AI are the same. Understanding the differences between various types of artificial intelligence is essential for evaluating their capabilities, limitations, and risks. Treating all AI systems as equally powerful can lead organizations, governments, and individuals to make poor decisions.

The book also explores why companies and institutions repeatedly fall for AI hype, why artificial intelligence cannot simply solve the problems of social media, and why the most immediate dangers of AI may come from the people and organizations deploying it rather than from autonomous machines acting on their own.

Instead of focusing primarily on speculative existential threats, the authors argue that society should pay closer attention to the real-world harms caused by algorithmic decision-making, unreliable AI systems, misleading marketing, lack of accountability, and the concentration of AI power in large technology companies.

AI Snake Oil also warns about a future in which increasingly powerful AI systems remain controlled by major technology companies with limited public oversight or accountability. The book emphasizes the importance of understanding who builds AI systems, who benefits from them, and who bears the consequences when they fail.

By examining the limitations of artificial intelligence, algorithmic bias, AI misinformation, unreliable predictive systems, AI governance, and responsible AI adoption, the book helps readers develop a more critical and realistic understanding of modern AI.

Whether you are evaluating AI tools for your business, using AI at work, making technology policy decisions, or simply trying to understand the rapid growth of artificial intelligence, AI Snake Oil provides a practical framework for separating real AI capabilities from marketing hype and bogus AI claims.