Saturday, April 9, 2016

The Fear of AlphaGod Is Not the Beginning of Wisdom

When “AlphaGo,” Google’s new super-smart game-playing computer dominated Lee Sedol, Korea’s so-called “Go Master” in five games of the ancient Chinese board game “Go,” doomsayers felt that they have approached the end-game: It is time for humans to cash in their chips –– the era of artificial intelligence has arrived.   
Go is an abstract strategy board game in which two players attempt to capture more territory than the opponent. The game seems simple enough when explained that way, but it was one of the only board games that had been considered unbeatable by computers. The number of possible moves in go easily surpasses the number of atoms in the universe, so winning by brute force, the endless trial and error method used by programs to eventually find a solution (rathe than using some ntellectual strategy, was and still is not a feasible tactic. To everyone’s surprise, however, AlphaGo systematically dismantled the 18-time world champion in a 4-1 victory. Although Lee narrowly clinched a win in Game Four, the forethought, the unpredictability, and the ease that AlphaGo demonstrated in the other four games revealed that AI technology has drastically improved and humans have a lot to learn from the program.
What scares a lot of people about AlphaGo is the very fact that it did not employ brute force to beat a Go master. People understand and accept that computers perform basic functions, such as calculation and information recall, better than human beings. However, in order to become the best at Go, one needs creativity and innovation––the traits that are uniquely associated with the human race. In Game Two, AlphaGo displayed something that can only be considered as creative, innovative, and even beautiful. Its move 37 was a highly unorthodox maneuver that everyone assumed to be a mistake, but over 100 turns later, it became apparent that it was the winning move.
By the means of machine learning and tree search techniques, AlphaGo learns the opponent’s style by watching how he plays and collecting relevant data. These pattern recognition techniques allow the program to dynamically predict his next move and to simultaneously calculate the best maneuver it could execute when facing that particular approach. This is not something humans don't do. By contrast, this is mimicking a distinct human trait, which becomes the basis for the excitement of thinking about its endless possibilities and applications as well as the fear that robots will eventually replace humans. Moreover, AlphaGo self-trains and self-improves by playing against other instances of itself, leading some to conjure up the harrowing picture of robots creating other robots to take over society.
The rise of artificial intelligence technology also raises a lot of profound philosophical questions. If artificial intelligence can operate at the same level as a human being, what does it mean to be truly human? And how do we define God, if we are able to create beings that can think, rationalize, innovate, and function better than we do? Are we breaching the realm of God and stealing the title for ourselves, or do we worship the almighty AIs that we created as our new gods?
Such existential questions, coupled with the natural fear of the unknown, which science zealously explores, understandably plant seeds of doubt and fear into people’s minds. Indeed, the fear of technology and science is nothing new. It pervaded history and manifested itself in literature as well as popular culture. Mary Shelley wrote Frankenstein to show that pursuing knowledge beyond accepted human limits is extremely dangerous. People in the 19th century enthusiastically agreed with Shelley, as the novel became an immediate best seller. Hollywood has again and again explored this topic with great commercial success, one of them being I, Robot­­––a movie about a dystopian society ruled by robots. The film was inspired by Issac Asimov’s short-story collection of the same name, which reflects the concerns and fears of the atomic age. Each time, people believed that it was game over for them –– a new technology would surely destroy the world soon.
Fortunately, humans emerged victorious, and the game continued. Even today, the doubts and fears about new technology neither provide sufficient grounds for nervously anticipating our doomsday nor present a need to kneel and pray to the AI gods that we will never reach the technological singularity. And we certainly must not delay or halt the development of artificial intelligence and robotics technology. By contrast, we need to embrace and promote such technologies for the betterment of society and human race. After all, those past technological advances that people feared have significantly improved and benefited our society. Likewise, the field of artificial intelligence will not replace human life; the technology exists to extend our longevity and enhance our quality of life.
The deep-learning technology AlphaGo employs has the potential to revolutionize everything we touch and use. Its ability to learn and predict its user’s behavior and preference by utilizing machine learning techniques and deep neural networks are crucial to improving our society in many respects. These methodologies are already being used as key components of inventing self-driving cars. More efficient data processing would allow us to categorize the Internet in a way that will end the days of hour-long Google searches to get the results that only moderately resemble what you wanted. Domestic robots equipped with advanced AI will flawlessly assess and fulfill their owners’ needs based on pre-programming and newly gained experiences on the job, greatly benefiting the elderly and disabled population. Medical robots will be able to perform difficult surgery with pinpoint accuracy and precision, and they will only become more reliable after multiple uses. With a certain amount of luck, it might one day even be able to make DMV offices operate properly.

With such useful applications and more, it becomes difficult to justify holding an aversion to the advent of AI technology based on the irrational fear that Siri might one day ask you for directions to AlphaGo’s house (although she would already know it). I, for one, welcome our new AI overlords since they would undoubtedly improve almost every aspect of our lives.

1 comment:

  1. Even though AI cannot replace the hands of a surgeon, or battlefield technicians at this moment, the prospects of what AIs can do is becoming more and more promising. AlphaGo has proven that the Artificial Intelligence research and development is rapidly advancing. This machine has beat the world’s best Go player Lee Sedol, by winning four games out of five, which is indeed the first time a machine has ever beaten its most skilled human counterpart. And just like what you said, the prospects are bright. Even though many people fear that AIs might take over the world, I believe that its development can significantly improve the quality of our lives. People once believed that the earth was flat and feared that traveling too far into the horizon would make them fall out of the planet. Some people feared that telephones would invade people’s privacy and criticized Alexander Graham Bell for creating the instrument of the Devil. TVs were thought to emit dangerous X-rays even after such problem had been long resolved. We also though Wi-Fis would emit invisible radiation that would cause terrible health effects. But everything only turned out to be significantly beneficial to our lives. I am excited to see what the advancing AI technology has in store for us.

    ReplyDelete