Experts see benefits of AI outweighing risks

Matt McDevitt, left, McKinsey partner and co-leader of QuantumBlack Japan, McKinsey AI Company, moderates a panel discussion during the Korea Times Forum at the KCCI Building in Seoul, Wednesday. From left are McDevitt; Kim Myoung-shin, LG AI Research principal policy officer; Chang Joong-ho, professor at Seoul School of Integrated Sciences & Technologies (aSSIST); Uhm Yul, director general of the Artificial Intelligence Policy Bureau at the Ministry of Science and ICT; and Yang Yoon-seok, assistant professor in the Department of Computer Science at the State University of New York, Korea. Korea Times photo by Choi Won-suk

The rapid growth of the artificial intelligence (AI) industry in recent years has reminded some people of themes from dystopian sci-fi movies, such as “The Terminator,” causing fear that machines may take over the world.

However, AI experts from corporate, academic and government sectors unanimously expressed an optimistic outlook for the future during a panel discussion at the Korea Times Forum, Wednesday, expecting the technology to have positive impacts for society.

Kim Myoung-shin, LG AI Research principal policy officer, emphasized that AI will help to overcome various crises facing humanity.

“AI needs a massive amount of energy at this moment, but the technological development has enabled consumption of less resources for better performance,” he said. “AI also predicts how drought, flood and wildfire will spread, so it contributes to dealing with climate change and natural disasters.”

Although Kim admitted that the gaps between countries and companies could be widened depending on whether they secure advanced technologies, he added that AI can alleviate geographic polarization to some degree by promoting the development of novel drugs and agricultural production.

Kim Myoung-shin, LG AI Research principal policy officer / Korea Times photo by Choi Won-suk

Chang Joong-ho, professor at Seoul School of Integrated Sciences & Technologies (aSSIST), forecasted the emergence of new jobs and industries that can counteract negative and illegal uses of AI.

“Numerous crimes and accidents have occurred in the wake of the growth of the internet and mobile businesses, and these problems have created new businesses, such as cybersecurity companies,” he said.

Uhm Yul, director general of the Artificial Intelligence Policy Bureau at the Ministry of Science and ICT, dismissed concerns over the famous saying, “AI will not replace you, but the person using AI will.”

“People using AI may reduce the number of jobs in the short run, but once the technology is applied to various industries, jobs will increase ultimately,” he said.

Chang Joong-ho, professor of Seoul School of Integrated Sciences & Technologies (aSSIST) / Korea Times photo by Choi Won-suk

Singularity already here

Yang Yoon-seok, assistant professor in the Department of Computer Science at the State University of New York, Korea, advised policymakers and companies to focus more on minimizing the negative impacts of AI, saying it is impossible to go against the trend of technological development.

He noted that the speed of AI’s development has been much faster than engineers expected.

“Ray Kurzweil defined singularity as the moment when artificial intelligence exceeds human intelligence, and a recent paper from Microsoft claimed that singularity is already here technologically,” he said. “I personally think that singularity is near.”

Yang Yoon-seok, assistant professor in the Department of Computer Science at the State University of New York, Korea / Korea Times photo by Choi Won-suk

Kim and Uhm agreed that singularity will occur in the near future.

“The happening of singularity relies on choice,” the LG AI researcher said. “If the government does not control the technological development, singularity may happen earlier than expected.”

Chang, however, pointed out that the AI industry will face regulatory hurdles someday, although he also believes that singularity will occur after society breaks through those barriers.

“The first AI winter in 1960s and the second winter in late 1990s resulted from technological problems,” he said. “The third winter will come, due to problems regarding ethics and regulations.”

Uhm Yul, director general of the Artificial Intelligence Policy Bureau at the Ministry of Science and ICT / Korea Times photo by Choi Won-suk

Cautious approach to AI regulations

There was a consensus among the panelists about the necessity of a code of AI ethics, although they maintained a cautious stance on regulating the technology.

As a policymaker, Uhm saw that regulations may have negative impacts on the industry, although the legally binding force could make companies act more socially responsible in their development of AI.

“The government has made efforts since 2019 to establish AI ethics, based on the OECD AI Principles,” he said.

Kim agreed to punishing companies that caused problems with their AI, but in terms of preventing the possible problems caused by AI, he recommended that the government induce investments to flow into ethical technologies, rather than prioritizing regulations.

“Users should also have AI literacy, so that they can avoid unethical companies,” he said.

Kim, who previously worked for UNESCO to draw up AI principles, pointed out that it could be difficult for different countries to follow a single global AI code of ethics, although most nations feel the necessity of such ethics. He therefore called for international discussions on which ethical values can be changed more flexibly.

“Each country has different ethical standards and beliefs, due to their different histories and traditions,” he said. “The biggest tasks in the AI industry over the next several years will be the agreement among the major players on the AI ethics.”

Chang noted the need for countermeasures against fake news generated with AI.

In addition, the professor who also runs a small AI startup asked the government to be clearer about regulations, so that smaller firms do not suffer from uncertainties.

Yang said AI companies and the government should prevent machine learning from using biased data, because biased responses from AI will have significantly negative impacts on students.

“We should be more careful about using AI at schools, and machine learning should be done with refined data,” he said. “Companies and the government should come up with better guidelines.”

Regarding this issue, Kim answered that LG has avoided web crawling in teaching its AI, saying that the company has secured millions of papers and theses it calls "qualified" in collaboration with its partner firm, to minimize AI hallucinations, confident responses by an AI that do not seem to be justified by the data a system is trained on.