Artificial Intelligence and the Wisdom of Physicists

I have been conditions to worry whenever physicists weigh in on the topic of artificial intelligence, particularly when they are Nobel laureates. It has now been 25 years since Roger Penrose published The Emperor's New Mind, in which he skillfully applied higher mathematics to demonstrate the impossibility of artificial intelligence. When a colleague equally skilled in the mathematics of computation theory pointed out the many flaws in Penrose's argument, all he did was write another book, Shadows of the Mind, which added little to the first, citing the flaws that had been raised but not addressing them with any substance.

Now we have an article by Stephen Hawking (with assistance from Stuart Russell, Max Tegmark, and Frank Wilczek), whose credentials are as good as Penrose's, arguing that artificial intelligence is inevitable and that we need to be prepared for consequences we might otherwise regret. Had Hawking been inspired by something other than the movie Transcendence, I might treat this article with more respect. After all, I agree with the punch line. I continue to believe that we have become so enamored of technology evangelism that we have all but expunged the word "consequences" from our working vocabulary. Furthermore, while his expertise is physics, Hawking certainly knows a thing or two about mind and embodiment that most of us will never adequately grasp. Nevertheless, like Penrose he also may not really grasp the reality of the technology, as opposed to the fantasies that are spun for the sake of funding research projects and startups.

So it goes.