Artificial intelligence and the future of the Korean economy

By Troy Stangarone

Manufacturing has historically been one of the keys to Korea's economic success. Korean firms have moved up the value chain to become world-class manufacturers in high-value sectors such as semiconductors, automobiles, consumer electronics and shipbuilding. However, that development has come at the expense of the Korean services sector. In an age where artificial intelligence (AI) and new innovations such as ChatGPT are emerging, an underdeveloped services sector could hinder Korea's future economic growth.

According to the World Bank, manufacturing accounts for about 25 percent of Korea's GDP. In contrast, manufacturing accounts for only 17 percent of global GDP and just 11 percent for the United States. Even in Germany, another highly industrialized economy, manufacturing accounts for only 19 percent of GDP.

While Korea has seen some successes on the digital side of the economy with Naver and Kakao, neither makes the Forbes 2000 list of the largest global companies. In contrast, Samsung is the world's 14th-largest company but has yet to add a strong digital or software component to its manufacturing prowess, in spite of its potential as one of the world's largest data companies due to its success in smartphones and efforts to be a leader in the Internet of Things.

Despite how it is often presented, AI is not new. The earliest example of AI is believed to be a checkers program developed by Christopher Strachey of the University of Oxford in 1952. Since those early days, AI has been used to provide basic tasks we take for granted such as predictive text in emails and text messages, personalized feeds on social media and recommendations on Netflix.

Each of these uses of AI has helped to either improve productivity, as with predictive text, or manage tasks that could not be handled at scale by humans alone, as is the case with social media feeds.

But new breakthroughs in artificial intelligence, such as those signaled by Microsoft's uses of OpenAI's ChatGPT in its Bing search engine, hold the ability to do much more. ChatGPT is able to not only search the internet more efficiently than existing search engines, but write reports on the topics it searches as well or better than humans. It can also generate code for new software programs.

These tools still need to be refined. Early reports on Microsoft's Bing integration suggest that the AI can move towards manipulation if pushed by users in certain directions, and still need to be checked by human users for accuracy. But on the whole, it can significantly reduce workloads.

Progress from AI may also not proceed as quickly as expected. The struggles with autonomous vehicles being one example. However, the technology will increasingly underpin not just customer-facing services, but research for the development and improvement of products, as well as be integrated into traditional manufacturing products to provide services. The integration of voice assistants into automobiles is one example, but also the prospects for AI integration into more common items such as refrigerators to monitor and order food.

Korea's success in manufacturing, specifically in semiconductors, provides one advantage in for the development of part of the AI ecosystem, but its failure to develop a world leading digital services and software industry could impair its manufacturing economy as AI is increasingly integrated into manufactured products and used to improve existing products and develop new ones.

At the moment, Korean firms are in a competition with Chinese firms to become the dominant providers of EV batteries to the automotive industry. While the Inflation Reduction Act provides Korean firms an advantage in the United States, Ford is looking for ways to circumvent the law's restrictions on Chinese EV batteries. In the EU, there are no restrictions on Chinese batteries. AI could help Korean firms develop better ways to produce batteries at lower costs and with longer ranges than their Chinese competitors.

The nature of AI means that some of the tools necessary for advances in AI can be adopted from abroad and do not need to be developed domestically. Startups and other firms can access cloud computing for the development of new AI applications, for example. But to take advantage of these tools Korea will need to foster development of a broader software and data services industry that will underpin advanced technologies related to artificial intelligence and artificial intelligence enabled services.

Some of this shift has already begun. The Moon administration adopted Korea's first AI strategy, and President Yoon Suk Yeol has encouraged the government to adopt new technologies akin to ChatGPT and pledged to make Korea the third-most-competitive nation in AI. Both SK Telecom and Kakao are planning to launch their own version of ChatGPT later this year.

However, for Korea to maintain its manufacturing edge and to reap the benefits of AI, it will need to develop a more robust software industry with access to high-quality data. Some of this can be achieved through the purchase of software companies abroad, but will also require support for entrepreneurship to foster new ideas and firms. AI will be the future, the question is whether Korea can adapt its own economic model quickly enough.

Troy Stangarone (ts@keia.org) is the senior director of congressional affairs and trade at the Korea Economic Institute.