Professor Kim Jin-ho from the Seoul School of Integrated Sciences & Technologies said the development of "deep learning," which refers to computers expanding their intelligence capacities by themselves, makes it complicated to predict the outcome.

AlphaGo, a program developed by Google, defeated European Go champion Fan Hui in October. Although the South Korean champion Lee Se-dol boasts far higher skills, some experts expect the program's capabilities to have also improved over the previous few months.

Industry watchers said AlphaGo can learn around a million game records in less than a month, which would take 1,000 years for a typical person. Google currently expects a 50:50 chance for either party to win.

Others expect that although AlphaGo has studied more than 30 million game records, it cannot defeat Lee in terms of judgment and intuition.

"While AlphaGo is excellent in making short-term strategies, it lacks skills in making decisions steps ahead," said Jeong Jae-seung, a professor of brain engineering at the Korea Advanced Institute of Science and Technology. "Lee is likely to win as the game of baduk calls for looking far ahead."

Some experts, however, said that even if Lee wins, the supremacy of mankind will not last for long.

"The victory or loss of AlphaGo in this game does not have a significant meaning," said Kim Dae-shik, another professor at KAIST. "AlphaGo will become stronger through the match with Lee. Estimating the outcome of another game years later will be really difficult." (Yonhap)