Korea faces long road in AI drive

Lack of IT originality a major obstacle to machine learning industry

By Lee Min-hyung

Korea’s tendency to mimic foreign technologies is a major obstacle facing the burgeoning local artificial intelligence (AI) industry, experts said Wednesday.

The AI industry is drawing public attention as machine learning is making headlines here thanks to a historic five-game go match, starting today between Google-developed AI system AlphaGo and world go champion Lee Se-dol.

Global IT giants such as Google, Microsoft and IBM, identify AI and machine learning as their next growth area, as the technology allows machines to make humanlike predictions by using massive datasets.

Local tech firms are also jumping on the AI bandwagon, but have failed to measure up to industry-leading standards as in the United States or Japan.

This reflects the nation’s decades-long habit of imitating foreign technology trends, said a computer science professor.

“Local technology experts tend to focus on studies which gain traction in developed countries,” said Kwon Hyuk-chul, a professor at Pusan National University, in an interview. “The nation was in the grip of big data fever five to six years ago, for the same reason. AI fever is also on the same track.”

He pointed out that the local AI industry cannot grow fast enough to meet global standards, unless the nations let go of such “me-too” attitudes.

“Experts and researchers focus too much on studying such areas as deep learning, machine learning and AI, all of which have gained overwhelming attention in recent years,” he said. “This attitude kills creativity, doing no good for the development of the nation’s actual AI industry.”

He added that government regulations are another obstacle to further AI research.

“Even with good intentions, the government does not give the green light to research which needs to pile up datasets,” said the professor. “I also understand that the government is in a serious dilemma over data management issues, but it sometimes has to ease regulations, at least for leadership in the promising AI industry.”

“Facebook would not have achieved such a global sensation if it had been developed in Korea,” he said. “The company would have been involved in lawsuits due to the nation’s complicated regulations. The government should urgently come up with a concrete measure to balance technological development and private data protection.”

Choi Kye-young, a senior researcher at the Korea Information Society Development Institute (KISDI), said in his recent report that local investments in AI have failed to catch up with the global standard.

“The AI industry accounts for only 3 percent of some 600 billion won investment allocated to the software industry here over the past five years,” he said.

The nation’s AI technology index stood at 75 points in 2014 under the benchmark 100 of the U.S., according to a report by the Institute for Information & Communications Technology Promotion (IITP) published in January 2015. Korea also lagged behind the 89.3 of Japan and average 89.8 for European countries, according to the report.

It added that Korea also suffers a two-year technology gap in the area, compared to the U.S.

Industry-wise, things are no different, as only a few companies here, such as SK Telecom and Naver, are engaging in AI research.

In particular, SK Telecom has been making investments in AI since 2012, incorporating the technology into its smart home services. The company is under negotiations with Apple and Google to jointly develop an AI platform.

Naver has also been running its own R&D arm Naver Labs since 2013 in its bid to lead the local machine learning industry.

However, most IT giants here have kept low profiles, as they cannot collect enough big data due to regulations on information sharing that block their investment appetite for machine learning.