In 2014 the philosopher Nick Bostrom published “Superintelligence: Paths, Dangers, Strategies” with Oxford University Press. The book asks how a machine intelligence far exceeding human capability might come about, what it would want, and whether humans could keep control of it. Bostrom argues that a sufficiently capable system could be difficult to constrain once created, and devotes much of the book to strategies for ensuring such a system’s goals would benefit humanity.
According to Bostrom’s own website, the book became a New York Times bestseller and “helped spark a global conversation about the future of AI.” His site lists endorsements from Bill Gates and from Sam Altman, who later co-founded and led OpenAI. The book gave a vocabulary - terms like the control problem, instrumental convergence, and the treacherous turn - that recurs throughout later AI-safety discussion.
Bostrom was Professor and founding Director of the Future of Humanity Institute at Oxford University from 2005 to 2024. Superintelligence built on earlier ideas, including I.J. Good’s 1965 intelligence-explosion argument, and packaged them for a general readership at the moment deep learning was beginning its rapid rise.
The Oxford University Press product page (ISBN 9780199678112) is the canonical publisher record; it did not return text to automated fetching during verification, so the book’s details here are drawn from Bostrom’s own site, which is a Tier 1 author source.