Since the birth of computer science, man has been asking that question: Can machines really think? Some noteworthy philosophers have argued that artificial intelligence is impossible (Dreyfus), that it is immoral (Weizenbaum) and that the very concept of it is incoherent (Searle).
Yet six decades ago, the father of computer science, Alan Mathison Turing, posed the Turing challenge. It stated that the age of machine intelligence would come when we could not discern between human and machine intelligence.
Since then, the world has wrestled with various cognitive models mimicking human intelligence. In 1966, Weizenbaum created Eliza to replicate the behavior of a Rogerian psychotherapist, which sometimes fooled people into thinking they were talking to a real person, by using rules that transformed users’ questions.
From the early Eliza and chatterbot modules to the more recent chess-playing Deep Blue, the world has started to wake up to the idea of machine intelligence. Today, we know computers can beat Jeopardy human champions (IBM’s Watson), cars can drive themselves (Google) and machines can follow rudimentary commands (Apple’s Siri). But rather than just domain specific game-playing or office management kinds of tools, the question remains: Can machines graduate to really emulate and rival human intelligence?
Francis Crick, the Nobel Prize-winning father of modern genetics and discoverer of DNA helix structures, used to opine that there is a fundamental framework of ideas that are missing to be able to interpret approaches to achieving machine intelligence. One thing is clear: If we are to clone human intelligence in all its generic thinking and problem-solving grandeur, we cannot fake it. We need to sincerely emulate the human brain. We need to study hierarchical temporal memory systems to gain insight into the theoretical neuroscience behind how human brains work.
Too often, we are tempted to take the course of studying a specific body of knowledge and combating combinatorial explosion by throwing computing power to distil copious amounts of knowledge into supercomputers. We ignore a pivotal suggestion from Turing, that the scalable way to make machines think is not to simulate the adult mind, but to simulate a child’s brain and then let it rapidly learn about the environment in which it finds itself. Adaptive learning is the key to unlocking the secrets of machine intelligence and fostering its ability to rival human intelligence.
Leveraging theoretical computer science principles including those taught by Turing, we are precipitously close to being the first to sincerely answer the six decade old Turing challenge. The idea would not be to just fake human behaviour to win the Loebner prize, but to make a sincere emulation of human brain that is capable of adaptively learning just the way a child learns, and rapidly becoming smarter and smarter by its interactions with humans.
What impact will thinking machines have on modern times?
It is hard to tell all the ramifications of machines starting to learn and think. When a prodigious child is born, it is hard to tell of the impact he will have when he grows up. What we do know is that we are on the precipice of a transformation unlike any before. When microprocessors were invented, they were predominantly developed for calculators and traffic light controllers. Today, the world is a more efficient shrinking village through the use of Internet and mobile communications.
As opposed to any schools of thoughts that preach “beware of machines,” we believe machine intelligence will lead the optimal form of creative destruction. Take a look at the world today. We are enslaved. The Pareto principle holds. 80% of the time we are caught in the trap of doing the same 20% of canonical, yet mundane, chores. Whether it is vacuuming the floor or driving a car, we are currently slaves to ordinary chores. Machine intelligence will serve as the ultimate liberator. It will liberate mankind to engage in higher forms of creative expression.
Will machines steal your jobs? Let’s ask ourselves what happened to the horse and carriage drivers. The invention of mechanized transport moved them up to driving cars today. People who worked with hand in car factories have now moved up to do computer-aided design and modelling of next generation cars. Necessity is the mother of invention. Machines will be the most faithful servants that will goad mankind to move brains higher up in the value chain. A human brain is a beautiful thing to waste.
Technological singularity, where machines dominate human intelligence, continues to be an elusive goal. If such capabilities were to be developed, then indeed it would yield an event horizon, beyond which the future of who’s ruling who would be hard to discern. But this Vigne concept is mere science fiction currently. Complex decision making, active reasoning and emotional intelligence are domains where human intelligence holds sway.
As much as science fiction authors prefer dystopian futures over utopian ones, so far, machine intelligence parallels, if not supersedes, other revolutionary technologies, whose negative repercussions are outweighed by their positive aspects. One thing we know for sure is that these are the most exciting times of transformation.