Do Not Adopt AI Out of Anxiety: First Know What You Need

Reflections on the course my boss asked me to attend last week — 4 Do not be anxious. Know what you need. Business owners are bombarded with information: AI will change the world, replace jobs, and leave companies that fail to transform digitally waiting to die. It sounds as though failing to upgrade today means closing tomorrow. At every dinner, their own table or the next is talking about AI, AI, AI. Anxiety mounts until they return to the company and announce that AI adoption starts now. But adoption for what? The answer is ‘reduce costs and improve efficiency’—another vague concept. What, specifically, will be done? The biggest problem is not knowing why. Following digital-transformation or AI trends alone leads to difficult data preparation, process changes, changes in people’s behavior, and disappointing benefits. Before long, the company gives up and returns to its old ways. The boss must understand the company, its problems, where to begin, and the problem’s boundaries. My greatest concern is that many Taiwanese bosses do not know, or cannot explain, what they want. Without self-understanding, problems and requirements keep shifting, making work difficult for internal teams, outside advisers, and startup teams. ---- This returns to my earlier posts. Taiwanese education is good at solving known problems but not at asking questions. Students spend days and weeks doing exercises, exams, memorization, and rote learning. They are not encouraged to spend time making projects, trying things themselves, or organizing a small team to discover or solve an interesting nearby problem. A Taiwanese student may complete over 10,000 exam papers, while an American may well undertake 100 projects or studies, discovering, defining, finding methods, and solving problems large and small. From childhood, Taiwan should encourage more questions: why do this, why set a specification this way, and what principle lies behind it? We excel at receiving specifications and working hard to meet them at lower cost. But without understanding the underlying reasons, we do not help shape specifications and can easily be replaced by lower-cost countries. ---- To conclude: do not be anxious. First establish what you need, the difficulties you will face, the investment required, and reasonable expectations.