Thanks for the invitation, @Sunflower Student.
Stop getting carried away by the celebration of AI-driven cost savings and efficiency.
A dangerous arrogance is taking hold in technology and capital. In venture-capital meeting rooms, people enthusiastically calculate how much labor AI will replace. On earnings calls, they boast about margins improved by layoffs. Meanwhile, the anguish of workers remains largely unseen.
This is not simply business-model innovation. It is a social liability whose reckoning has been deferred.
That arrogance can dangerously distort strategic investment. If business leaders focus only on return on investment and ignore the possibility of sweeping policy responses, today’s spending may eventually become a burden their companies cannot bear.
1. The corporate illusion: a fatal omission from the ROI calculation
Many corporate AI strategies share a feature: their ROI calculations treat labor substitution as the sole source of profit.
The capital-side calculation is simple: invest 100 million in AI research, replace 200 workers and save 30 million in labor costs. On paper, the deal looks compelling.
My judgment is that this calculation is incomplete from the start.
A company survives through social as well as technical relationships. When many companies simultaneously compress labor income and new jobs, other income sources and transfers fail to keep pace, consumer demand can weaken. An individual company may sell to other groups or overseas; one layoff does not inevitably cost it its customers. But the economy-wide transmission of income belongs in its long-term calculations.
When unemployment pressures reach a critical level and replacement costs become costs of cushioning social disruption, celebrated efficiency gains can turn into substantial liabilities.
2. The broader perspective: governments intervene as economic eras change
This strategic blind spot has entered international policy debates.The IMF’s 2024 analysis of AI and fiscal policyexamined how social protection, training, tax incentives and taxation of capital income might address potential employment shocks and inequality. It did not present a dedicated AI tax as an automatic answer. Instead, it emphasized that such a tax could discourage investment and be difficult to implement.
The lesson for companies is to include institutional responses to employment and distribution in scenario planning. Countries may choose different combinations of protection, training and taxation. Businesses should assess these separately rather than assume that today’s rules will remain unchanged.
Historically, governments have responded to disruptions in production through labor protections, taxation and social insurance. These experiences can inform thinking about AI, but cannot predict every institutional response:
The Industrial Revolution: the initial expansion of steam-powered production was accompanied by extreme hardship for workers. Governments eventually intervened through Factory Acts and labor protections, redirecting some gains toward workers and easing structural social divisions.
The age of mass production: assembly lines improved efficiency in parts of manufacturing, but expanding output also required demand backed by purchasing power. Wages, labor relations, social protection, credit and taxation jointly shaped consumption. The emergence of the middle class cannot be attributed to one institution, nor does workers’ inability to buy a particular product necessarily mean economic collapse.
Efficiency gains from technological progress do not belong to companies alone. When their distribution becomes unbalanced, governments may respond through taxation, social protection, education and labor institutions. The timing, scale and tools depend on political choices and fiscal capacity; there is no single inevitable path.
3. A strategic warning: do not turn technology into a weapon against society
Employment protection, social insurance and taxation may all adjust to automation. Corporate strategy must take that possibility seriously. Betting everything on purely labor-replacing AI is also an advance commitment to substantial future policy risk.
As an observer of robotics, I suggest considering the following:
Redefine return on investment: ROI should include more than payroll savings. Consider how inclusive the business is. Improving safety in dangerous settings or enabling employment for disabled and disadvantaged people may attract public support and customer recognition. Tax treatment still depends on the actual rules; social value does not automatically confer a tax exemption.
Create value through human–machine cooperation: consider the evolution of leading robotics companies as they move from robots as replacement tools toward robots as extensions of physical capability. This is a constructive direction for compliance, social acceptance and long-term development.
Engage with rules rather than resisting them: thoughtful companies participate early in shaping standards. Can you help establish industry standards for upgrading workers’ AI-related skills? Can some gains be distributed through supply-chain contracts, creating roles as maintainers of AI rather than leaving people displaced by it?
4. In this period of rebalancing, are you building assets or accumulating liabilities?
Operators who assume AI brings only profits and no costs may discover that wages saved are not the whole return. Transition support, supervision and maintenance, shifting demand and future tax changes can all alter the outcome. Whether these costs arise, and how large they are, must be estimated separately.
In this new cycle of rebalancing, use technological progress as a foundation for shared gains with society. This is more than a moral choice. In 2026, it is a practical survival strategy for businesses navigating the AI era.
If you still insist that only total labor replacement counts as innovation, recognize what is at stake. This is more than an operating question; it concerns a company’s political and social judgment, and potentially its survival. Businesses that ignore social costs and see only financial statements risk being excluded by a government-driven rebalancing.
Remember: no company is so large that it can place itself above social stability.
The author’s original article appeared on Toutiao: “A Strategic Warning in the AI Wave: Do Not Let Technological Arrogance Blind Corporate Direction.”