Divination Engines

“Dazzling in its archival depth and urgent in its implications, Divination Engines is a landmark in the critical history of artificial intelligence. Li brilliantly argues that the statistical turn in natural language processing was never just a technical choice. It was an epistemic wager, staking statistical prediction against the harder work of understanding. The consequences of that bet now surround us, from chatbot hallucinations to AI slop. This is a book for anyone who wants to understand how language became data and what we lost in the translation.”

“Have you ever used Siri or ChatGPT and wondered: How did these systems come to be? And why are they desirable? Divination Engines answers these questions through a brilliant and engaging analysis of the transformation of speech recognition technology from hardware-based signal processing devices to data-driven natural language processing systems. Through this history, Li reveals how human experience in general became subject to predictive and generative algorithms. A tour de force that explains why and how these systems came to matter.”

“A detailed, closely researched history of not only the way that language became computational but also the current impulse toward the datafication of everything, Divination Engines shows how knowledge itself is being transformed. Ambitious in its culminating argument and empirically grounded, it will be an important book in the history of big data and AI.”

Xiaochang Li is an assistant professor in the Department of Communication at Stanford University and an affiliate faculty member in the Program in Science, Technology, and Society and the Program in Modern Thought and Literature.