Our increasingly smart machines are now being applied to activities that not long ago were viewed as the exclusive domain of humans. Cognitive computing, robotics, self-driving cars and AI in general are some of the hottest topics in academia and business these days.
What should we expect from this new generation of AI machines and applications? Are they basically the next generation of sophisticated tools enhancing our human capabilities, as was previously the case with electricity, cars, airplanes, computers and the Internet? Or are they radically different from our previous tools because they embody something as fundamentally human as intelligence?
Kevin Kelly, - as am I, - is firmly in the AI-as-a-tool camp, as evident by a recent article in Wired, - The Three Breakthroughs That Have Finally Unleashed AI on the World. Kelly co-founded Wired in 1993, where he served as executive editor until 1999 and continues his association as Editor-At-Large. He has many other interests as well as being a prolific writer of books and articles.
According to Kelly, the AI future that’s coming into view is nothing like the “potentially homicidal” HAL 9000 from 2001: A Space Odyssey or “a Singularitan rapture of superintelligence.” The AI he foresees is more like a kind of “cheap, reliable, industrial-grade digital smartness running behind everything, and almost invisible except when it blinks off. This common utility will serve you as much IQ as you want but no more than you need. Like all utilities, AI will be supremely boring, even as it transforms the Internet, the global economy, and civilization.”
He compares AI to electricity. “Everything that we formerly electrified we will now cognitize.” Like any other tool, this “utilitarian AI will also augment us individually as people (deepening our memory, speeding our recognition) and collectively as a species. There is almost nothing we can think of that cannot be made new, different, or interesting by infusing it with some extra IQ. In fact, the business plans of the next 10,000 startups are easy to forecast: Take X and add AI. This is a big deal, and now it’s here.”
Artificial Intelligence was one of the most exciting parts of computer sciences in the 1960s and 1970s. Many of the its early leaders were convinced that computers would as intelligent as a human being within a generation. In the 1980s, the Japanese government even mounted a major national program, the Fifth Generation Computer Project, to develop advanced AI machines and programming languages. But, after years of unfulfilled promises, a so called AI winter of reduced interest and funding set in everywhere.
Several years later, a different AI paradigm emerged. Instead of trying to program computers to act intelligently, - an approach that hadn’t worked because we don’t really know what intelligence is, - AI now embraced a statistical, brute force approach based on analyzing vast amounts of information with powerful computers and sophisticated algorithms.
Such a statistical, information-based approach produced something akin to intelligence or insight. Moreover, unlike the earlier programming-based projects, the new paradigm scaled very nicely. The more data you had, the more powerful the supercomputers, the more sophisticated the algorithms, the better the AI results. Deep Blue, IBM’s chess playing supercomputer, demonstrated the power of this new paradigm by beating then reigning chess champion Garry Kasparov in a celebrated match in May of 1997.
In his article, Kelly lists the three key breakthroughs that have finally unleashed the long awaited, “60-years-in-the-making” success of AI:
“Cheap parallel computation: Thinking is an inherently parallel process, billions of neurons firing simultaneously to create synchronous waves of cortical computation. To build a neural network - the primary architecture of AI software - also requires many different processes to take place simultaneously.”
“Big data: Every intelligence has to be taught. A human brain, which is genetically primed to categorize things, still needs to see a dozen examples before it can distinguish between cats and dogs. That's even more true for artificial minds. Even the best-programmed computer has to play at least a thousand games of chess before it gets good.”
“Better algorithms: Digital neural nets were invented in the 1950s, but it took decades for computer scientists to learn how to tame the astronomically huge combinatorial relationships between a million - or 100 million - neurons. The key was to organize neural nets into stacked layers… In 2006, Geoff Hinton, then at the University of Toronto, made a key tweak to this method, which he dubbed deep learning. He was able to mathematically optimize results from each layer so that the learning accumulated faster as it proceeded up the stack of layers. Deep-learning algorithms accelerated enormously a few years later when they were ported to GPUs.”
Given that in the foreseeable future we should expect even more powerful parallel computing, lots more data, and deeper algorithms, AI will keep improving and become an increasingly important part of everyday life.
Deep Blue’s 1997 victory over Kasparov was a major AI milestone. But having shown that machines could beat chess grand masters, the research experiment was over and people lost interest in such contests. Kasparov himself realized that he could play better chess if he had the same access to the data base of all previous chess moves that Deep Blue had, and he pioneered the concept of advanced chess, in which human centaurs use computer tools to augment their chess playing capabilities as they compete against other such man-plus-machine centaurs. Intagrand, a team composed of various humans and chess programs, won the 2014 Freestyle Chess tournament and is considered today’s best overall chess player.
“But here’s the even more surprising part: The advent of AI didn’t diminish the performance of purely human chess players,” writes Kelly. “Quite the opposite. Cheap, supersmart chess programs inspired more people than ever to play chess, at more tournaments than ever, and the players got better than ever. There are more than twice as many grand masters now as there were when Deep Blue first beat Kasparov. The top-ranked human chess player today, Magnus Carlsen, trained with AIs and has been deemed the most computer-like of all human chess players. He also has the highest human grand master rating of all time. If AI can help humans become better chess players, it stands to reason that it can help us become better pilots, better doctors, better judges, better teachers.”
Kelly nicely explains the difference between human intelligence, - the result of millions of years of evolution, - and the data-driven smartness of AI. Like with any other tool, AI applications will be highly specialized, doing what they do really, really well but little else. “In the next 10 years, 99 percent of the artificial intelligence that you will interact with, directly or indirectly, will be nerdily autistic, supersmart specialists.”
Human intelligence, on the other hand, includes qualities like self-awareness, introspection and self-consciousness. It’s quite likely that such qualities spark our creativity and help us come up with ideas never explored before. Why not attempt to infuse our machines with such human attributes so that they too can become more creative?
This is indeed a fascinating research question. But, such human intelligence would be a serious liability in a real world AI machine. “We want our self-driving car to be inhumanly focused on the road, not obsessing over an argument it had with the garage. The synthetic Dr. Watson at our hospital should be maniacal in its work, never wondering whether it should have majored in English instead. As AIs develop, we might have to engineer ways to prevent consciousness in them - and our most premium AI services will likely be advertised as consciousness-free.”
He wonders if rather than artificial intelligence, what we want instead from our machines is something less grandiose but more focused, measurable and specific - artificial smartness. When it comes to physical qualities like power and speed no one expects our machines, - e.g., electric motors or cars, - to in any way have to imitate humans. Similarly, the smartness of AI machines need not be similar to our human intelligence.
In fact, Kelly believes that one of the chief virtues of AI is its alien nature compared to human intelligence, because it can then better complement our own capabilities. In addition, “it will help us better understand what we mean by intelligence in the first place. In the past, we would have said only a superintelligent AI could drive a car, or beat a human at Jeopardy! or chess. But once AI did each of those things, we considered that achievement obviously mechanical and hardly worth the label of true intelligence. Every success in AI redefines it.”
“But we haven’t just been redefining what we mean by AI - we’ve been redefining what it means to be human. Over the past 60 years, as mechanical processes have replicated behaviors and talents we thought were unique to humans, we’ve had to change our minds about what sets us apart… In the grandest irony of all, the greatest benefit of an everyday, utilitarian AI will not be increased productivity or an economics of abundance or a new way of doing science - although all those will happen. The greatest benefit of the arrival of artificial intelligence is that AIs will help define humanity. We need AIs to tell us who we are.”