The Technological Singularity (MIT Press Essential Knowledge series)
Like many others who have dedicated their working lives to research in artificial intelligence, I was inspired as a child by science fiction. My boyhood hero was not a real person. It was Susan Calvin, the scientist in Asimov’s I Robot stories (the written works, not the film) who pioneered the field of robot psychology. More than anyone else, I wanted to be like her when I grew up. Now that I have (sort of) grown up, and in real life bear the title of Professor of Cognitive Robotics, I have a more complex relationship with science fiction. I still see it as a source of inspiration and as a medium for exploring important philosophical ideas. However, the ideas it explores merit a deeper treatment. The primary purpose of science fiction is to entertain, albeit in an intellectually stimulating way. It would be a mistake to use it as a guide to thinking.
So this is not intended as a work of science fiction. Nor is it a piece of so-called futurology. The aim here is not to make predictions. Rather, it is to investigate a range of possible future scenarios, without committing to the prospect of any one in particular, and without any particular timescale in mind. Indeed even highly unlikely or remote scenarios are sometimes worthy of study. This is true, for instance, if a scenario is especially dystopian. In that case we might want to think carefully about how to reduce its likelihood even further. Unlikely or remote scenarios are also worth discussing if they raise interesting philosophical questions, obliging us, for example, to think about what we really want as a species. So whether or not you think we will soon create human-level artificial intelligence, whether or not you think the singularity is near, the very idea deserves some serious thought.
This is a short book on a very large theme. So it can only stand as an introduction, with many important issues given only a brief treatment. For example, various arguments relating to consciousness are presented to which there are well-known counterarguments, and these merit counterarguments of their own. But an introductory book has to skip over these subtleties. Also the focus is heavily on the future of artificial intelligence, and some significant related topics, such as nanotechnology and biotechnology, are barely touched on. The book is intended to provide a neutral overview of the conceptual territory, and I have attempted to outline both sides of the argument in controversial matters. However, it seems unavoidable that some of my own views will be visible through the veil of neutrality, despite my best efforts.
I would like to thank the very many people who have discussed artificial intelligence with me over the decades, not only academics and students but also members of the public who have attended my talks. I would like to thank them all by name, but that would be impossible. So I will reserve my explicit gratitude for a few colleagues whose recent influence has been especially pertinent. Thanks to Stuart Armstrong, Nick Bostrom, Andrew Davison, Daniel Dewey, Randal Koene, Richard Newcombe, Owen Holland, Huw Price, Stuart Russell, Anders Sandberg, and Jaan Tallinn. Sorry to those I have forgotten. Finally, I would like to thank MIT Press, and especially Bob Prior, for encouraging me to write the book in the first place.
North Norfolk and South Kensington, October 2014
|August 28, 2018
How to Read and Open File Type for PC ?