Ecosystem

How Strong is AI Development in South Korea?

The future is AI Technology. Top entrepreneurs like Elon Musk and Mark Zuckerberg have been talking about the potential value and harm of true AI. AI represents a major leap forward in how computers can learn independently. In South Korea, a vast majority of jobs are jobs that require little creativity and a lot of manual repetition. These jobs include retail, customer support, and factory jobs. Currently, many computers can perform a particular task over and over again. Therefore they can learn as they go. However, is this really AI? Or rather, is it just machine learning? It is difficult to find companies in Korea that are creating true AI, however, most are moving in that direction.

AI vs. Machine Learning

Many startups in Korea are branding themselves as AI technology startups. However, in most cases, they are working towards AI by using Machine Learning. Therefore, Many have not achieved true AI. True AI means some software that can solve problems by itself without any help from humans. Think of AI as the ultimate goal, and machine learning is a subset of AI. Machine learning is one of many tools, including deep learning and reinforcement learning. Machine learning uses large amounts of data to find the best solution. Therefore they build algorithms that can learn from data. These are designed to perform specific tasks only.

We saw with Google’s Deep Mind that a computer could get smarter by performing many simulated games (data) repeatedly. This is how the world’s top GO player Lee Sedol lost to a computer (AlphaGo). A self-driving car learns through its owner as well as the environment. Therefore these self-learning machines can become better drivers over time. As AI technology advances, how far is Korea when it comes to this breakthrough technology? How much effect will AI have on the Korean economy and, more importantly, employment in Korea?

AI in Korea

So far, one of the top Korean AI systems is called Exobrain. It was developed by the Electronics and Telecommunications Research Institute (ETRI). Exobrain, a few years back, beat two quiz champions and two collegians. One got a perfect score on Korea’s national college entrance exam. Not only did Exobrain beat the human counterparts but crushed them by getting a score of 510. To take that into context, the next highest score was 360. Exobrain won the top prize of 25 million won.

“Exobrain’s victory today is a milestone that marked a new era in Korea’s AI research,” said Seo Suk-jin, director general of the Software Policy Bureau of the Ministry of Science, ICT, and Future Planning. “We will concentrate on supporting development of the Exobrain AI system, which will be the key in the fourth industrial revolution, and help it compete with other AI systems such as IBM’s Watson in the industry.”

Exobrain was not perfect, as it still got a few questions wrong. Those questions tended to be more subjective rather than objective. This will be a relatively easy fix in the future as more data is put into the system. This will help in its ability to analyze and learn.

So what can Exobrain do? Exobrain can analyze the Korean language, learn from it, and constantly improve its analysis. Therefore Exobrain can store large amounts of data which it will use to answer even the most complex questions. Therefore this is more of a machine learning system than true AI, but the potential is intriguing.

Big Corporations in Korea Developing AI Technologies

Naver

Naver (the Google of Korea) and KAIST (the MIT of Korea) are partnering up to create an AI research center that focuses on the application of hyperscale computing. What is hyperscale computing? It is the processing of big data and cloud computing. Therefore hyperscale AI refers to the analysis of massive datasets to make autonomous decisions. Major global companies like Amazon, Google, and Microsoft are already using this technique. The center will be called “KAIST-NAVER Hypercreaive AI Center”. It will have over 100 AI researchers from both Naver and KAIST to develop hyperscale AI and creative AI. Naver has also partnered with Seoul National University (the Harvard of Korea) to join the project.

KT

KT has aggressively embraced AI technology in Korea by investing and researching in AI/Machine Learning fields. Their focus now is on AI, big data, and cloud computing. Like what AWS is doing for Amazon, KT will look to Cloud services to expand its business beyond voice communication and high-speed internet services. Therefore, AI will be critical in developing cloud services that easily sync with Korea’s 5G network. As 5G increases in Korea, the need for cloud services in Korea will also increase.

All-in-One Services

KT will focus on bringing an all-in-one service that includes network, storage, and wireless hardware. Therefore they will look to continue their research into their AI speaker GiGA Genie. Currently, it is being utilized for hotels, cars, and homes. However, they want to expand their service for customer service, robots, and automation for the food and beverage industry. Therefore, KT has been aggressively looking to partner with AI startups in Korea and conglomerates in the AI field to expand their AI network. Recently they partnered with Dongwon Group to create an AI Smart Factory that will package Dongwon’s fishery products. In addition, KT partnered with Hyundai Heavy Industries Holding, KAIST, Hanyang University, and Korea Electronics and Telecommunications Research Institute to establish an AI One Team. They will focus on creating an Open AI ecosystem to further the growth of AI technology in Korea.

Hyperscale AI in Korea

KT is also looking into hyperscale AI through joint research with Korean Universities and government research centers. They partnered with AI One Team, a research alliance, debuted the hyperscale AI model in 2022. AI One Team includes ten organizations, with KAIST being one of them. KT was in charge of the whole R&D process, mainly data collection. The Universities KT has partnered with will be responsible for machine-learning algorithms and data filtering. Training of the AI model finished in 2021, and the commercialization of the AI platform will start in the first half of 2023. The hyperscale AI will be first applied to KT’s AI assistant Giga Genie.

“Developing a hyperscale AI model which combines world-class technologies and experiences of AI One FTeam has special meaning in securing core technology in the next generation of AI. Through this project, we will contribute to strengthening the competitiveness of Korean AI,” said Hong Kyung-pyo, an official from KT.

Kakao Enterprise

Kakao Enterprise is Kakao’s AI technology-based working platform. As far as creating partnerships, Kakao has been very aggressive. They aim to become the AI department store in Korea. They have partnered with KBS to create an AI announcer and Kyobyo Life Insurance to create an AI chatbot for their customer support center. They are also working with Samsung C& to build a service that allows Everland (Disneyland of Korea) ticket sales and food/beverage orders using KakaoTalk. Moreover, they are working with LG Electronics to connect Kakao’s AI speaker, Kakao Mini, to all their TV models.

Kakao Brain

Kakao Brain is Kakao’s AI research subsidiary. Their large-scale AI models can understand the context and meaning of sentences without the need for human assistance. It has been trained to understand what sentences mean and make new sentences or images from what they have understood. Their focus for 2022 will be to create an AI program that can analyze and understand different data input modes, such as text, image, and video. The aim is to get everyone their own AI teacher and doctor.

Samsung Electronics

To dominate the global AI industry, a company must create semiconductors with AI abilities. Samsung will look to create AI semiconductors by 2030. The Korean government fully supports helping the next generation of semiconductor companies. They plan to invest over $850 million in AI chips by 2030. In addition, Samsung has also stated that they will use AI to create new ships using Synopsys AI software when designing their ships.

South Korea already has two of the world’s biggest memory chip makers: Samsung and SK Hynix. However, this is more on the manufacturing than the core technology side. Samsung’s focus will be on making semiconductors smarter using AI technology. Samsung has pledged to hire over 1,000 new employees to work on AI technology.

South Korea’s Artificial Intelligence Research Institute

The Artificial Intelligence Research Institute (AIRI) was established by the Software Policy & Research Institute (SPRi) and brought some of the top companies in Korea together. Some companies included Samsung Electronics, LG Electronics, SK Telecom, KT, Hyundai Motor Company, and Hanwha Life Insurance which put 3 billion won into the project back in 2016. Since then, the AIRI has been used to research and develop intelligence information technology. This does include not only AI but also big data and cloud technologies.

Self Driving Cars in Korea

President of South Korea, Moon Jae-in wants self-driving cars in Korea by 2027. His goal is to become the first country in the world to have self-driving cars on major streets. The Korean government and Hyundai Motor Group plan to invest over $50 billion in the self-driving car industry by 2025.

“Our goal is to become the world’s number one powerhouse in the future vehicle industry by 2030,” said Moon Jae In.

The Korean government has already started to revise regulations and set up traffic infrastructure for fully autonomous vehicles in Korea. In addition, they will draw up rules and regulations for insurance regarding self-driving cars. The goal for Hyundai is to have fully autonomous cars on the streets of Korea by 2024.

Korean Startup Using Self-Driving Technology

Korean startup RideFlux is looking to build a ride-sharing company. They have partnered with another mobility startup, SOCAR, to create Korea’s first self-driving car. The self-driving car has cameras and proximity sensors and has been running on Korea’s roads for a while now. They have already transported customers from Jeju airport to local tourist destinations. Under South Korean law, a safety driver has to sit behind the wheel, poised to intervene if necessary.

“Our aim is to put self-driving cars without safety drivers operating in all parts of Jeju in the next five years,” said the CEO of RideFlux Park Jung-hee.

The Korean Government’s Support for AI Technology

The Korean government is heavily investing in AI technology. They released an AI strategy that would put them as a top 5 contender in AI by 2023. The Korean government realized that they needed more AI engineers to make this a reality. They have already made plans to build 6 new AI schools. This will lead to more than 5,000 AI engineers by 2024. In addition, the Korean government will focus on AI-based industrial intelligence across all manufacturing sectors. They have created an AI strategy that will lead to the development of over 2,000 AI-based factories by 2030. These will be more advanced versions of smart factories.

The data collected from these factories will be stored in a data center to support AI-based services in software, sensors, robots, and equipment. The Korean government plans to invest over $7 billion in R&D in three main sectors: non-memory chips, biotech, and mobility. A little under $10 billion was spent from 2020-2022 in these sectors. The rest was used for data, 5G network services, and AI. They hope these investments in core infrastructure and new industries will help innovative growth that will spill over into other industry sectors. One of the first tasks will be to launch an AI research and development complex in Gwangju.

Korea to become the most competitive AI industry.

South Korea dominates the semiconductor industry and has 25 times the processing speed of existing integrated circuits. Korea will continue to invest billions into the next generation of smart chips. Therefore Korea will be at the cutting edge of semiconductor technologies moving forward. An example of the latest cutting-edge semiconductor technology is processing-in-memory (PIM) chips, which can act like a human brain. Over 1 trillion won will be used for processing-in-memory chips that can mimic a person’s brain and facilitate “deep learning” by machines that are critical for the expansion of AI not only in Korea but around the globe.

The aim to become an AI industry leader is strong as the Korean government feels this AI technology will bring in hundreds of billions of dollars in the long term. In addition, AI technology can potentially improve the standard of living in Korea. South Korea’s AI technology is behind the US, China, and the European Union. However, Korea hopes to narrow this gap greatly by 2030. The first step towards this is to support Korean AI startups. Furthermore, the Korean government has promised to invest over $20 billion of its budget to nurture the science, technology, and AI sectors.

Korea is already one of the world’s leaders in manufacturing and 5G communication network infrastructure. It is only a matter of time before they have a top 5 global AI infrastructure.

John

John is the Co-Founder of Seoulz. He has covered the Korean startup & tech scene for over eight years and has written over 700 articles regarding the Korean startup ecosystem. He has brought global attention to Korea's tech scene using Google SEO. Email him at john@seoulz.com

Recent Posts

Korea‘s Deep Tech Moment: Investors and Startups Ride the Wave – An Interview with BlissVine Ventures’ CEO

Deep Tech is quickly becoming a key focus in the global investment landscape, and it…

4 months ago

Top Fabless Semiconductor Startups in Korea

As machine learning and AI technology continue to advance, businesses all around the world require…

5 months ago

AR in Korea – The Potential to offer Better Customer Experiences

By now, many of us have already been exposed to the world of augmented reality.…

7 months ago

Mycovation harnesses Mycelium to produce Better food

As the world struggles with food security and nutrition challenges, there is growing acknowledgment that…

7 months ago

HerbiaEra – Substrates that Simplify and Optimize Indoor Growing

Undoubtedly, agriculture is one of the most fundamental industries on our planet. It provides food,…

8 months ago

Inspektlabs – A Video-Based Guidance App for Car Inspections

As anyone who owns a car knows, the current vehicle inspection process is not only…

8 months ago