[Viewpoint] Beware the Pitfalls of Adopting AI Technology

Kim Kyung-jun, Vice Chairman of Deloitte Consulting

At the center of the megatrend where the flow led by digital technology companies is spreading to analog companies lies AI. AI, which was once perceived as science fiction until the 2016 Lee Sedol vs. AlphaGo Go match, has quickly become a common term. Companies are also actively adopting AI. They are driving innovation by integrating AI into the entire process from procurement to production, sales, and research and development.

Interest in AI is an encouraging phenomenon, but excessive enthusiasm inevitably leads to bubbles. A typical example is coding education for all employees promoted by some companies. It only increases organizational fatigue while delivering low effectiveness relative to the input. Just as programmers who code may not understand the business significance of AI, general employees who spend time and effort learning coding are far from actual application. Similarly, the expectation that AI will instantly solve long-standing challenges is unrealistic. This is a typical bubble created by so-called experts who package AI as a panacea for impatient companies. AI technology, at the level of general companies’ utilization, is close to a general-purpose technology. It is sufficient to purchase and use the necessary parts in the required areas.

In the 1970s mainframe era, computer experts were mysterious figures akin to priests in a temple. Although computers were introduced as cutting-edge devices in science magazines and newspapers, few people actually encountered them. Large computers, used only in limited fields such as advanced weaponry and space exploration, were difficult for the general public to even see. However, nowadays, ordinary people use computer technology for their purposes without understanding it. Teenagers use it as a gaming device, students in their twenties as a learning tool, and workers in their thirties for work processing. It would be convenient to understand computer technology, but it is not necessary. It is enough to turn on the computer and run applications.

Consider a logistics company whose main business is land transportation. It moves goods via trucks through logistics warehouses. The logistics company does not need to know the base and manufacturing technologies of trucks. It is sufficient to have the ability to maintain and manage the core tool, the truck, so that it is always operational. Minor repairs are handled internally, and major repairs are outsourced to external repair shops. Trucks that have reached the end of their lifespan are processed at external scrapyards, and replacement trucks are purchased from manufacturers to successfully operate the business. All other tools, such as hammers and pencils, only require knowledge of how to use them according to their purpose. Detailed technical knowledge related to tools is nice to have but not essential.

AI technology, too, is sufficient for general companies to use as a tool according to their purposes. AI technology development and utilization capabilities are separate. For AI technology-specialized companies, the technology itself is a competitive advantage, but the position of general companies is different. Because AI technology is rapidly advancing, it is difficult for general companies to keep up. General companies only need to secure the ability to apply commercialized AI technology based on domain knowledge in their business areas. Domain experts who have acquired basic AI technology are actually more suitable for deriving optimal solutions.

When introducing new technology, one must be cautious of the ‘technology trap’ of becoming obsessed with the technology itself. The same applies to AI. For general companies, AI is a generalized tool used to solve problems. The core of innovation using AI is not the technology but the domain knowledge of the field. In AI adoption by general companies, an effective approach is to equip internal personnel who have accumulated business experience, i.e., domain knowledge, with the basic competencies needed to utilize AI technology and to collaborate with external AI technology experts.

[Kim Kyung-jun, Vice Chairman of Deloitte Consulting]

This content was produced with the assistance of AI translation services.