The AI economy
I consulted a navigation app to travel to Songdo in Incheon. It advised me to avoid the expressway, which gets crowded on weekends. As soon as I got out of Seoul, the roads were jammed. As I blindly followed the slow-moving traffic, I arrived at a traffic light. The left-turn signal was so ridiculously short. When the signal was green, it stayed that way for a long time. Frustrated cars ignored the signal and made left turns, causing mayhem at the intersection.
I passed a similar intersection on my way back. The traffic there was smooth. The difference was a notice on the signal board. It said that left turns or U-turns were perfectly okay if done with caution. All the cars - going straight, turning left, wanting to make U-turns - were able to move. A road sign succeeded at a task that an expensive navigation system flubbed. Leaving matters to human judgment and flexibility solved a traffic bottleneck. Flashing signals late at night and the first-in-first-out principle in intersections also work well. The effect is good, the cost is low.
Rigid order and regulation often gets in the way of efficiency. Much of Korea and the global technology community had been engrossed with a historic match between a machine and a tournament master in the ancient board game of Go, called baduk in Korean. No computer was able to conquer the game until Google’s artificial intelligence algorithm AlphaGo. World Go champion Lee Se-dol was humbled with a single win in a five-game tournament that ended on Tuesday. Lee has a 9-dan rank.
But the match between 33-year-old Lee, whose humility and calm shone throughout the nail-biting matches against the strongest-ever, rapidly evolving software that had the help of a vast neural network of computer connections, made an indelible impression on the Korean public. Koreans across generations have now become familiar with the term artificial intelligence, or AI. Scientific words like algorithms and advance progress in AI-based products and technology now pop up in daily conversations.
The government did not miss the significance of the event. The Ministry of Science, ICT and Future Planning announced plans to increase investment in AI following the match. It pledged 30 billion won ($25.7 million) in AI research and development. The minister toured start-ups to check on Korea’s status in the AI field. The minister of trade, industry and energy also held a meeting with industrialists and experts to discuss applications of AI technology and innovations. The Trade Ministry pledged to significantly increase the budget for AI research separately from the Science Ministry.
This is, of course, all good news. Korea’s AI industry supposedly lags 2.6 years behind the United States and even China. It is a laggard at every level, from basic research to commercialization. AI is expected to revolutionize not only the way industries work but also the very lives of mankind. Regardless of Lee’s defeat, the AlphaGo phenomenon could be a windfall for Korea. The AI field, which requires patient and farsighted investment, desperately needs government support.
But people on the research and industrial side don’t seem particularly thrilled. State support often means government nagging and meddling. The government has a track record of those unpleasant habits. Under government prodding, Korean scientists were asked to deliver a local alternative to the computer operating system MS-DOS in the 1990s. They turned out K-Tube six months after YouTube went viral. The results were disastrous. Few know the existence of these research projects, not to mention any success resulting from them. It was because the state demanded fast results that nothing worthwhile resulted. Bureaucracy kills creativity in basic scientific research.
The five top universities - Seoul National University, Korea University, Yonsei University, Korea Advanced Institute of Science and Technology, and Pohang University of Science and Technology - joined forces to protest the government’s support system, which prioritizes quantitative outcomes. AI is the most advanced sector that incorporates information communications and digital technology. To catch up with advanced nations, we must first establish an efficient research and development infrastructure. A new path must not be ruined with ill-placed traffic signals.
Directions and speed must not be regulated. The fewer signals there are, the better the path will be for AI development. The road is best left to the judgment of drivers. They must be free to make turns and find shortcuts on their own. The government deregulation initiative that allows everything except for some basic rules should be first employed on AI research and development.
JoongAng Ilbo, March 17, Page 34
*The author is an editorial writer for the JoongAng Ilbo.