Price Is Decided Not by Theory, but by Real-World Tensions
Price is determined by supply and demand. That is how economics explains it. But on the actual front line where companies decide prices, the judgment is grittier. What are competitors charging? What is the cost? Are labor costs rising? What about logistics costs? How is the exchange rate moving? How much should we protect the gross margin? If we raise the price, will customers leave?
In other words, while price is theoretically determined by supply and demand, in practice it is determined by a tug-of-war among cost, competition, customer tolerance, and profit margin.
From this premise, the impact of AI on prices becomes quite complex.
Even When Costs Fall, Companies Don't Necessarily Lower Prices
If AI advances automation and efficiency, corporate costs go down. Back office, customer support, translation, research, document preparation, ad operations, e-commerce operations, parts of development. Much of the work that humans once spent time on can be done in a fraction of the time with AI.
So if costs fall, will prices fall? In theory, they can. Even if demand stays constant, if supply-side costs fall, companies can offer goods and services at lower prices. In a highly competitive market, one company cuts prices and others follow, and the whole industry's prices come down.
But reality is not that simple. Companies do not necessarily lower prices just because costs have fallen. In markets with weak competition, strong brands, or high switching costs, the cost savings remain with the company as improved profit margins, not as lower prices for consumers. AI is more likely to be used first as a corporate profit-improvement device, not as a consumer rebate device.
Especially in Japan, this view is probably closer to reality. For many Japanese companies, the goal of adopting AI is not "lowering prices." It is making up for labor shortages, standardizing personalized work, reducing overtime, covering the shortage of young workers, suppressing outsourcing costs, and improving decision-making accuracy. In other words, AI is used less to lower prices than to maintain and improve business operations.
The Costs AI Lowers Are Not the Costs Driving Up Today's Prices
What matters here is that the costs AI lowers and the costs currently pushing up prices are not necessarily the same.
What AI tends to lower is the cost of knowledge work, clerical work, routine tasks, and information processing. Meanwhile, what is currently pushing up prices in Japan is the weak yen, imported raw materials, energy, logistics costs, labor shortages, interest rates, and geopolitical risk.
Food, fuel, imported goods, rent, healthcare, elderly care, logistics. These will not easily get cheaper through AI alone. Even if domestic operations become somewhat more efficient, as long as import prices and energy prices stay high, final prices will stay elevated.
Therefore, even as AI advances, it is hard to expect overall prices to come down in a straightforward way. What is more likely to happen is a reorganization of the price structure itself.
What AI Can Copy Gets Cheaper; What It Cannot Gets More Expensive
What AI can easily replace gets cheaper. Research, translation, simple design, article writing, report preparation, first-pass analysis, routine customer support, simple development, parts of ad operations, parts of e-commerce operations. These are areas where, until now, price was attached to human time. Once human time is replaced by AI, prices come under downward pressure.
On the other hand, what AI cannot easily replace gets more expensive. Trusted specialists. Judgment backed by front-line experience. The ability to push complex projects forward. Brand. Place. Primary information. Real experience. Healthcare and elderly care. High-quality service. Scarce materials. Urban real estate. None of these increase easily just because AI does. If anything, the more information and content are generated cheaply and in bulk, the more the value of truly trustworthy people, places, and experiences rises.
AI does not lower all prices. It lowers the price of what AI can copy, and raises the price of what AI cannot copy. This is the essence of the polarization to come.
Work and Income Polarize the Same Way
The same applies to employment. AI will not take away all human jobs, but it will significantly restructure what those jobs contain. Jobs that produced value mainly through routine document preparation, coordination, search, summarization, or first-pass analysis face downward pricing pressure. Conversely, jobs that understand the front line, define the right questions, mobilize stakeholders, and carry the work through to implementation rise in value.
As a result, income may polarize too. Those who ride AI to raise productivity, and those who get replaced by AI. Companies that use AI to redesign their business, and companies that use it only to prolong existing operations. Companies that have capital, data, and customer touchpoints, and companies that used to take on simple, contracted work.
This gap widens over time.
What Japan Will See: A Distorted Price Structure
What about overall purchasing power? This is not monolithic either. If AI raises productivity, corporate profits increase, and those profits flow into wages and investment, real purchasing power improves. Even if prices do not fall, life gets easier as incomes rise.
But if profits are concentrated in a small number of companies and capital owners and do not flow sufficiently into worker wages, the purchasing power of the middle class and below declines. In that case, discretionary spending weakens and downward pressure on prices appears. Meanwhile, necessities and imports stay elevated. The result is a distorted state where "what is cheap gets even cheaper, but what life requires stays expensive."
What is likely to happen in Japan is, most probably, this form. A full-blown return to deflation is highly unlikely. Nor can it be called healthy inflation. AI lowers some service prices. Some companies see profit margins improve. White-collar jobs get restructured. Meanwhile, the weak yen, import costs, energy, logistics, and labor shortages keep necessities elevated.
In other words, what feels digital and intangible becomes cheap, fast, and convenient. But eating, living, moving, getting medical care, asking another person for help, consulting a trusted specialist — these areas do not get cheaper.
The Real Question: The Cheapening Side, or the Side That Grows More Valuable?
Price changes in the AI era are not a simple question of "will prices go up or down." Areas where price was attached to human time face price collapse from AI. Areas where price is attached to scarcity, trust, front-line judgment, brand, place, or real experience, by contrast, get more expensive.
What matters for companies is not just lowering costs with AI. It is figuring out whether your own products and services are on the side that AI cheapens, or on the side whose value rises in the AI era. If you are on the side that gets cheaper, operational efficiency alone is not enough. Before getting swept into price competition, you have to redesign the value you offer. If you are on the side that gets more expensive, you have to refine that scarcity and trust into something customers can actually feel.
AI is not just a cost-cutting tool. It changes the premise of pricing, the value of work, and the structure of corporate profits. And it changes what customers are willing to pay for.
That is why the question to ask in the AI era is not "how much cheaper can we make it." Are we on the side that AI makes cheaper? Or are we on the side that grows more valuable the more AI spreads?
That is the question from which the business has to be re-examined.
← Back to Thoughts