Atlas of AI

Power, Politics, and the Planetary Costs of Artificial Intelligence

Kate Crawford

13 min read
1m 30s intro

Brief summary

Atlas of AI reveals that artificial intelligence is not an abstract technology but a material industry built from minerals, energy, hidden human labor, and vast datasets. It argues that understanding AI requires tracing the physical and social costs behind its claims of autonomous intelligence.

Who it's for

This book is for anyone who wants to look beyond the hype and understand the real-world material, labor, and social costs of artificial intelligence.

Atlas of AI

Audio & text in the Readsome app

Introduction to Artificial Intelligence

The story of Clever Hans, a horse in the early 1900s that appeared to solve complex math problems, serves as a foundational lesson in how we perceive intelligence. While the public believed Hans possessed human-like reasoning, a psychologist named Oskar Pfungst discovered the horse was actually responding to subtle, unconscious physical cues from his trainer. This phenomenon, known as the Clever Hans Effect, illustrates a central trap in artificial intelligence. Systems that appear to perform spectacularly may simply be reflecting the biases and cues of their creators rather than possessing genuine understanding.

Artificial intelligence is often described through two misleading myths that obscure its true nature. The first myth suggests that non-human systems are direct analogues for the human mind and can be built from scratch without considering how humans are physical, social, and relational beings. The second myth treats intelligence as something natural and independent of social, cultural, and political forces. In reality, the concept of intelligence has historically been used to justify domination and inequality, and modern technology often inherits these historical prejudices.

Author Kate Crawford argues that artificial intelligence is neither truly artificial nor intelligent. Instead, it is a material industry built from natural resources, human labor, and massive infrastructures. To understand this technology, we must look beyond code and algorithms and view it as an extractive industry that takes heavily from the physical world. This perspective requires mapping the shifting landscapes of power to see who benefits from these systems and who bears the costs.

Ultimately, artificial intelligence functions as a registry of power that reinforces existing social hierarchies. It is a manifestation of highly organized capital backed by a small number of dominant corporations. By abstracting away the material conditions of its creation and extracting resources from those with the least power, the industry avoids accountability. Understanding artificial intelligence as a political and economic force allows us to make collective decisions about the role technology should play in our future.

Full summary available in the Readsome app

Get it on Google PlayDownload on the App Store

About the author

Kate Crawford

Kate Crawford is a leading researcher, writer, and academic who studies the social and political implications of artificial intelligence. As a Senior Principal Researcher at Microsoft Research, a Research Professor at USC Annenberg, and the co-founder of the AI Now Institute, her work focuses on the intersection of large-scale data systems, machine learning, and their historical, political, and environmental contexts. Crawford has advised policymakers globally and her collaborative art projects investigating AI have been featured in the permanent collections of the Museum of Modern Art and the Victoria and Albert Museum.

Similar book summaries