Settings

ⓕ font-size

  • -2
  • -1
  • 0
  • +1
  • +2

INTERVIEWKorea advised to avoid heavy-handed AI regulations

  • Facebook share button
  • Twitter share button
  • Kakao share button
  • Mail share button
  • Link share button
Visitors watch humanoid robots dancing at the World Artificial Intelligence Conference (WAIC) 2023 in Shanghai, China, July 6. Xinhua-Yonhap
Visitors watch humanoid robots dancing at the World Artificial Intelligence Conference (WAIC) 2023 in Shanghai, China, July 6. Xinhua-Yonhap

This is the second in a three-part series of interviews with global artificial intelligence (AI) technology experts regarding Korea's ambition to become a hub for AI innovation so that the country can reap AI's national security and economic benefits. _ ED.

'Seoul's strengths in chips to help US-China de-escalate tensions in AI'

By Kim Yoo-chul

Since the fanfare over generative artificial intelligence (AI) technology began this year, experts have voiced major concerns about the potential side effects of such technology.

Top experts explicitly warned about the risk of human extinction from the growth of AI, stressing that mitigating such a possibility must be a global priority.

In recent months, the U.S. and China have specified plans on how to better regulate AI technology. OpenAI CEO Sam Altman used his recent high-profile global trips to call for coordinated regulation by governments.

The EU was the first to enact legally-binding regulations that would apply to different areas of AI. However, the key concern is that many top business leaders in the EU were resisting rules on AI, warning that legislation could spur an exodus of investments and also jeopardize the region's technological sovereignty.

More than 160 big tech companies in Europe including Siemens, Renault, Airbus, Carrefour, Meta and ARM sent an open letter to EU lawmakers arguing that the EU AI Act goes too far, especially in terms of the regulation of generative AI and foundation models that are the technology behind today's mainstream AI platforms such as ChatGPT. EU member states plan to vote on the rules.

CAPTRS co-founder and CEO Phil Siegel / Courtesy of CAPTRS
CAPTRS co-founder and CEO Phil Siegel / Courtesy of CAPTRS
Speaking to The Korea Times, Phil Siegel, co-founder and CEO of the Center for Advanced Preparedness and Threat Response Simulation (CAPTRS), said the EU AI Act maybe leaning too much on regulations, while not paying enough attention to benefits.

"Korea seems balanced but it will all be in how it is implemented. Too heavy handed and rules will stifle progress in reaping benefits. Too light touch and rules could allow progress to run too far ahead of regulatory priorities," Siegel, who is also a commentator on AI for MSNBC and Fox Business, said in a recent interview.

CAPTRS is a U.S.-based nonprofit organization that uses AI to improve social disaster preparedness. Its key clients and partners include IBM Research, Boston Consulting Group, The New York Academy of Sciences, Center for Strategic and International Studies, and The Rockefeller Foundation.

When asked about the goals of regulations on AI technology, Siegel stressed it is more of a framework, which is good.

"I believe the four key areas for AI rules focus for any country or region should be protecting children, hardening and updating laws to recognize AI, ensuring fairness and accuracy in the algorithms and ensuring general safety and guidelines including registration and ethics."

The EU AI Act is mainly intended to propose organized responses to the risks AI is starting to pose as EU lawmakers want AI systems like ChatGPT to disclose the differentiations between original content and AI-generated content to ensure safeguards against illegal content. But big tech companies such as Meta, which owns Facebook and Instagram, brushed off potential side effects of AI by claiming that now is not the time to hit the pause button.

Tesla CEO Elon Musk speaks via video link at the opening ceremony of the World Artificial Intelligence Conference (WAIC) in Shanghai on July 6. Electric car giant Tesla is set to realize fully autonomous vehicles 'later this year,' in the words of CEO Elon Musk, in the billionaire's latest forecast for the long-anticipated milestone. AFP-Yonhap
Tesla CEO Elon Musk speaks via video link at the opening ceremony of the World Artificial Intelligence Conference (WAIC) in Shanghai on July 6. Electric car giant Tesla is set to realize fully autonomous vehicles 'later this year,' in the words of CEO Elon Musk, in the billionaire's latest forecast for the long-anticipated milestone. AFP-Yonhap

In Korea, the AI Act, which would take effect from this year at the earliest, is aimed at offering a statutory basis for establishing ethical guidelines for AI and providing support for new businesses in the industry.

"I think that is a practical prescription," Siegel said.

Korea's focus on education, guardrails make sense

Korea is in the early stage of detailing and specifying roadmaps for AI goals and regulatory guidelines from an economic perspective. However, Siegel said AI is well defined by Korea, which stipulates the technology as essentially "mimicking human intelligence derived using machines or digital formats."

"The goals and guidelines were both pragmatic though as with most new technologies will need to be fine-tuned. The focus on education is smart and it is really the only way for Korea to catch up on the software and algorithmic side of AI, though I must admit it feels like more of a long shot than the hardware and semiconductor side where it goes in competitively very well-positioned," according to the CEO.

As AI spreads across economies and societies, policymakers and stakeholders in the industry are seeking to move from principles to practice. Siegel said while it is very necessary for every country to have an overall technological strategy, perspective is what everyone will struggle with.

"I believe Europe is going maybe a bit far on regulation and not focused enough on benefits and skill training and reaping the benefits of AI. The U.S. is still figuring it out as are much of Asian countries. The U.S. advantage is that Silicon Valley won't wait. The disadvantage is they won't wait. I like Korea's focus on education and regulatory guardrails together as a balance. I think also enabling its chip companies to flourish will be strategically important," he added.

Competition between the U.S. and China is heating up and this is most evident in AI. From Washington's viewpoint, Beijing's wider use of AI for military purposes is another key source of concern. The core purpose of the U.S. decision to impose restrictions on its semiconductor exports to China is because of concerns that Beijing could possibly use sophisticated chips and AI to advance its military capabilities.

Nadine, third from left, one of the world's most realistic humanoid social robots (University of Geneva), Nadia Thalmann, fourth from left, from the University of Geneva, the head of an AI robot, of Hanson Robotics, female robot Mika, Ameca, right, one of the world's most lifelike humanoid robots, speak during the World's first press conference with a panel of AI-enabled humanoid social robots as part of the International Telecommunication Union (ITU) AI for Good Global Summit in Geneva, Switzerland, July 7. EPA-Yonhap
Nadine, third from left, one of the world's most realistic humanoid social robots (University of Geneva), Nadia Thalmann, fourth from left, from the University of Geneva, the head of an AI robot, of Hanson Robotics, female robot Mika, Ameca, right, one of the world's most lifelike humanoid robots, speak during the World's first press conference with a panel of AI-enabled humanoid social robots as part of the International Telecommunication Union (ITU) AI for Good Global Summit in Geneva, Switzerland, July 7. EPA-Yonhap

As a longer term strategy, Washington is weighing whether to impose additional curbs on investments in Chinese technology firms as China has long been dependent on U.S. investments to power its promising startups. Because the U.S. hopes to defend its advantage in managing advanced AI systems capable of completing complex cognitive tasks at a faster speed, Korea could increase its global influence in the growing AI industry.

Korea is home to the world's two largest memory chipmakers ―Samsung Electronics and SK hynix. Generative AI such as ChatGPT is a type of AI capable of creating content such as text, images, code and more. A significant number of memory chips are required to operate ultra-large language models as memory chips are used to read and write data. Samsung in March said it is planning to invest up to 300 trillion won in a new semiconductor facility, here.

"If Korea can grow its chip sector capabilities quickly, it will become even more strategically-important to the world and frankly help the world de-escalate tensions between China and the West. The Pharma model is a good one ― have Samsung and SK not only become strong but make sure large increases in startup investments in new chip companies are being made," Siegel stressed.

"Those companies could 'grow up' to be a number three but more likely get acquired by the big guys ― just like U.S. Pharma. The key is encouraging the entrepreneurial spirit and encouraging Samsung and SK to let engineers spin out to start new ventures," he added.

The government said last year that it would earmark 1.02 trillion won in funding for AI semiconductor research and development (R&D) over the next five years. Korea will also allocate 826.2 billion won through 2030 to help local companies manufacture advanced chips through new data centers and working with startups.



Kim Yoo-chul yckim@koreatimes.co.kr


X
CLOSE

Top 10 Stories

go top LETTER