In response to the latest film on the topic of artificial intelligence (AI) entitled, “Ex Machina,” it is interesting to revisit the topic of where we stand scientifically regarding whether/when the first computer will be constructed that can think on its own.
The expert who enables this discussion to be had in the first place belongs to Prof. Turing, the mathematician who paved the way for the world’s first full-fledged computer.
Ever since then, the notion has been out there that computers can threaten our job security. Dating back 66 years ago mathematician Norbert Weiner wrote the book entitled, “the Human Use of Human Beings” in which he predicts correctly the takeover of certain jobs by machines. First there was the machine built for assembly line menial labor, and now there is the machine built to take over many administrative jobs.
What will happen next? Will machines ever be made that would pass the so-called “Turing test” for intelligence, and pass as humans? Will machines take over all jobs? If so, what will this mean for humans?
Importantly, one must ask the question of whether AI will ever become a reality. Perhaps the world’s expert on the science behind AI is Emeritus Prof. Freeman Dyson, who works at the Institute for Advanced Study in Princeton, New Jersey. Prof. Dyson is one of the more humble founders of the physics field of quantum mechanics, and has also been thinking and writing on the topic of AI for the last 55 years.
His idea is that the understanding of our own brain will parallel our ability to develop AI. How can we create thinking machines that model our own brains if the understanding of cognitive function still so limited, and even naive? After all, we do not yet understand all the physical mechanisms involved with low-level cognitive functions, let alone the subtleties of emotions such as devotion to friends and family.
Dyson points out that one limitation is that computers are digital in form while the human brain is analog (although with some digital components). He goes on to argue that while digital AI may be possible, it is much more likely that the first AI will have an analog mind.
This is because part of the success of the human brain is our ability to use analog means to do certain types of comparative analysis that is still more accurate and faster than a computer. Perhaps driven by the question of whether AI is possible, and/or to solve other computational problems, we may find motivation to go back and advance decades-old analog technology. Analog or not, when will AI happen, if ever?
Prof. Dyson’s sums up the attempt to make an AI over the past 50 years as a dismal failure, and says such attempts will likely continue in that direction until we increase our knowledge of the brain. Using the past as a guide, he would wager that AI is at least 100 years away, but admits that he could be wrong. If it even happens at all, he envisions that such beings will not have a need to harm us as in the science fiction movies, or even to thank us, but simply to leave Earth to satisfy their growing need for energy to supply their significant artificial brains. But how will we feel about them?
