+ A

Lessons from Moravec’s paradox

May 28,2018
이미지뷰
Kim Hyoung-tae
*The author is a guest professor at Seoul National University Business School and former president of the Korea Capital Market Institute.

Moravec’s paradox refers to a discovery by artificial intelligence (AI) and robotics researchers in the 1980s, including Austrian-born Hans Moravec. Contrary to general belief, they found that high-level reasoning requires very little computation while low-level sensorimotor skills require enormous computational resources. Moravec wrote, “It is comparatively easy to make computers exhibit adult-level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a 1-year-old when it comes to perception and mobility.”

AI technology has evolved at a staggering pace. Machines won a world championship in the game of chess in 1997 and in the game of Go in 2016. AI-driven unmanned vehicles are expected to hit roads soon. Computers can give advice on stocks and treat medical patients. But they are still far from mastering a 5-year-old’s ability to perceive surroundings and build toy blocks.

Why is that? What is believed to be born natural to humans has evolved over millions of years. Human experiences and development have become fixated as a natural course in our consciousness. All the physical and mental movements needed for survival come naturally and easily because they are inputted in the brain and neuron systems.

What are complicated procedures to machines can easily be conceived by humans “at a glance” to be either understood or found strange. It is hard to expect insightful interpretation of human behavior and any kind of sharp sensibility from a machine. What makes a human has been evolved and polished over millions of years and comes naturally to humans. It may be nearly impossible for machines to replicate these processes. No matter how smart they are, they cannot beat an evolutionary process of millions of years.

So why does AI excel in such brainy games as chess and Go? The Western board game has a history of about 3,000 years and the Asian one 4,000 years. Compared to the evolution of mankind, that span of time is the wink of an eye. If the history of human evolution is compared to 24 hours, the history of Go would take up just five seconds. Due to the relatively short period of time, not all people find the games easy.

Such board games are an area in which computing can do better than humans. The human brain cannot compete with a computing process of algorithm of millions of past game plays. I am not underestimating AI capabilities. I am just saying the realm of difficulty and complexity differs for human and machines.

Both corporations and individuals fear the ascent of AI. To maintain human competitiveness in the age of automation and robotics, I suggest more research on human nature and its capabilities. More thorough study and understanding of human capacity can discover strengths that AIs cannot replace, as well as better robotic capabilities to supplement human weaknesses.

Appraisers do not examine fake works. They study real art in and out and build the insight to differentiate the real from the fake. What won’t change in the AI age are the authentic parts of human nature. The original has a history, but also incorporates the future. Discovery of the original can lead to discovery of an “old future.” The innovations of the future can be successful when they respect the primary features and functions.

The George Foreman Grill is a portable, electrically heated grill that has sold over 100 million units around the world. With a clamshell design the size of a laptop, the grill does not have any temperature control or a button to choose the type of meat. It was promoted by former boxing champion George Foreman. He is known to have earned millions of dollars from the use of his name, more than he earned as a professional boxer, which is a high-paying profession.

I did not have many expectations for an appliance taking its name from a heavyweight boxer. But I was surprised by the stunning taste of a steak grilled on it. It was juicier than steaks served in steakhouses. Moreover, it is easy to use — just put the steak on the pan and close the lid. There is no need to worry about the time, temperature or choice of meat. The non-stick grill can be cleaned with a wipe of a kitchen towel. Although simple in design, the machine has lasted and sold steadily since 1994 because it is true to the grill’s original function of cooking a steak well.

The design is simple and reliable like the boxer. As the boxer’s tagline on commercials says — “It’s so good, I put my name on it!” — the grill stuck to the basics of its function, which is to cook a good steak. The more complicated the world and advanced technologies become, the more people long for simpler products true to their original function.

Original products should differ from primitive products. The first has added high tech to the original function, but hides its advanced technology in a simple design. Businesses in the future will stay successful when they focus on products staying true to their original functions and remain respectful of human nature. Artificial intelligence cannot change what is innate. It should be an instrument to maintain and augment the features of the original. That’s another lesson from Moravec’s paradox.

Translation by the Korea JoongAng Daily staff.

JoongAng Sunday, May 26-27, Page 35