Even his own engineering and R&D teams only half understand artificial intelligence, says Shen Kuo-jung, chairman of GUS Technology and president of the Hota Group, who believes Taiwan's traditional manufacturers need at least a year or two before they can use it well. He made the case in an exclusive interview with Storm Media.
Shen leads GUS Technology, a Taoyuan-based maker of lithium battery cells that trades on the Taipei Exchange's emerging board. He stepped down as chairman of Hota Industrial Mfg. Co., a Taichung maker of gears and shafts, in August 2025 and handed the post to his daughter, according to Newtalk, while staying on as Hota Group president. In the interview, he described three hurdles facing traditional industry: talent, data and cost.
Seed Teams From Universities Anchor Shen's Plan
Shen's answer to the talent gap is to go back to the classroom. Companies, he said, can set up dedicated programs to build an in-house seed team, sending six to eight employees to study in relevant university departments with the company backing the training. Those employees would then help the rest of the organization understand and apply AI.
He named National Taipei University of Technology (Taipei Tech) and National Taiwan University of Science and Technology (Taiwan Tech) as examples of schools that could help through information, semiconductor or computing expertise. Shen was describing a direction rather than announcing a program: no dedicated class has been set up with any specific school.
"This is definitely not something you can learn in a few months," Shen said. Understanding how AI is trained, how it learns, how to talk to it and how to apply it will take "at the very least a year or two," he said, before traditional industry can gradually learn, transform and reform.
The estimate suggests he does not see AI adoption as a one-time purchase. Employees have to start from their existing expertise, build hands-on experience and carry the new skills back to the shop floor.
Data Protection And On-Premises Costs Weigh On Adoption
Money and security complicate the picture. Shen said some manufacturing information must be protected and cannot simply be handed to outside systems without an assessment. Running AI on a company's own servers, the alternative, can be relatively expensive.
That makes adoption more than a job for the technology department. Companies must decide how to protect their data, how much to invest and how to measure the payoff. Whether the data can be used at all, and whether a given solution meets the company's needs, will shape the next round of spending. For Shen, the choice between cloud and on-premises systems also involves how data is handled and who has access, not just how easy a tool is to use.
Veteran Craft Knowledge Must Become Teachable Content
Factory floors have long depended on the experience of veteran technicians, so the first question for a company is where to begin, Shen said. He used bread making as an illustration: judgments about ingredients, saltiness, sweetness and aroma could all offer openings for AI, but the real challenge is turning that experience into material a system can learn from.
Engineers' manufacturing expertise still has to connect with the new tools, he said. He suggested that teams with the relevant skills bring equipment to the factory and run drills, observation visits and internships, so that employees learn actual operation instead of treating a briefing as the end of the job.
Engineers Struggle Even With Basic AI Operation
The difficulty starts with the basics, Shen said. Entering data, getting a system to learn and conversing with it all require relearning. "The engineers and R&D teams inside our company really only half understand AI," he said.
Shen believes AI will raise productivity, and that traditional industries that stay out risk slower progress and weaker competitiveness. But agreeing that the technology matters and knowing how to adopt it are separate problems.
The question is pressing for companies in the middle of leadership succession, where the incoming generation must take over markets, finances and production management while also building the ability to use new tools.
Social Media Links & Editor Tag - Penny
