“If 'industrial AI', which increases added value by applying artificial intelligence (AI) to industrial sites, is activated, profits of more than 100 trillion won can be created only in the semiconductor industry.”

Kim Young-han, CEO of Gauss Labs, considered a world-class scholar in data science, said this at the 'AI Future Forum (AIFF) Webinar' held on the 24th. In order to advance this future, it is urgent to converge industry-specific expertise and AI technology, and some point out that the government should expand support for industrial AI.

“Industrial field-based AI technology needs to be developed”

The AI ​​Future Forum is the largest AI research network in Korea, where about 300 experts from industry and academia have gathered for the development of AI in Korea. Two discussion forums were held this March and last month, and the third webinar was held on the same day under the theme of 'AI expectations and reality, what kind of effort is needed'.

CEO Kim, in charge of the keynote speech, is a tenured professor at UC San Diego, USA, and has been leading Gauss Labs, a company specializing in AI research established by SK Group in Silicon Valley in the US since August last year. He argued that “AI is currently being used a lot in service fields such as AI assistants, but when it is applied to industrial sites such as manufacturing, it can make a much bigger impact.”

Citing a report from McKinsey, a global consulting firm, released in April this year, he said, “The potential benefit of using AI to the global semiconductor industry alone will exceed 100 trillion won.” McKinsey predicts that the profit that AI will contribute to the semiconductor industry will grow to $45 billion (about 50 trillion won) in the short term and to 85 billion dollars (about 100 trillion won) in the long term. CEO Kim explained, “If AI is properly used, the entire manufacturing process from semiconductor design to production and inspection can be optimized and enormous time and cost can be reduced.”

However, he points out that the reality has not yet met expectations.

CEO Kim pointed out, “There is a lot of data in the industrial field, but the ability to select high-quality data is low, and the development of AI algorithms and applications tailored to the characteristics of each industry is slow.”

To overcome this reality, he emphasized that AI development based on industrial sites is necessary. He said, “In the AI ​​industry, there is a tendency to develop AI tools first and then find out where they will be used, but the order is wrong.” It is necessary to first understand the nature and core challenges of a specific industry, and then develop AI tools accordingly.

In his opinion, it is also important to realize excessive expectations for AI. “AI is not like magic that solves all human problems,” said CEO Kim. .

AI created by participation… “Possible No. 1 in Korea”

Professor Jang Young-jae of the Department of Industrial and Systems Engineering at KAIST evaluated that the performance of manufacturing AI is gradually appearing. He said, “When we applied the AI ​​solution developed by KAIST to the field of intelligent factory operation, it showed much better productivity than engineers with 10 years of experience.

In the field of financial AI, it is analyzed that technology startups need to advance. Park Ki-eun, Executive Vice President of Kookmin Bank, said, “In addition to the already common chatbot and optical reading (OCR) fields, there are wide fields to expand deep learning technology such as investment prediction and algorithm trading. A structure should be created that forms an ecosystem with the

Jo Dong-yeon, a defense AI expert and professor of military affairs at Seogyeong University, advised, “Civil-military cooperation is important, and the scope of AI application should be expanded beyond the military field to respond to infectious diseases and natural disasters. ”

Kyung-Jeon Lee, a professor of business administration and big data applications at Kyunghee University, said, “If AI technology is developed from a thoroughly scientific and rational point of view, Korea can become the world's first in the AI ​​​field.” He continued, “We need to make sure that services such as the chatbot 'Iruda', which have caused controversy, can continue to operate.”

Reporter Seo Min-jun/Lee Si-eun morandol@hankyung.com

Source (in Korean): Hangkyung