The AI boom is global. So is its bill — Opinion & Editorial Columnists

TEXT START: The United States and the Philippines occupy different positions in the AI economy, but both must ensure that its gains are not built on unstable power systems, displaced workers and costs transferred to the public.

THE DISSECTION

This is an establishment transition memo disguised as accountability journalism. It accurately inventories AI’s physical and social costs—electricity demand, grid strain, public subsidies, labor displacement and unequal global gains—then reduces a structural rupture to a policy-management problem. Its central maneuver is to admit enough disruption to appear serious while preserving the belief that regulation, retraining, infrastructure investment and international cooperation can keep the existing economic order intact.

It describes the machinery around the guillotine, then recommends better queue management.

THE CORE FALLACY

The article treats the collapse of human economic necessity as a distributional problem. Under the Discontinuity Thesis, if AI achieves durable cost and performance superiority across cognitive work, the issue is not merely who receives the gains. The issue is that fewer people are required to produce the output that sustains the system.

“Job transformation rather than wholesale replacement” is a dangerously narrow frame. Replacement can occur through hiring freezes, vanished entry-level roles, compressed teams and declining bargaining power without a single spectacular mass layoff. Exposure statistics also understate the real threat: a worker can retain nominal employment while losing the wage, progression path and economic leverage that made the job viable.

Retraining does not recreate sufficient demand for displaced labor. Labor obligations and human-review requirements may slow deployment in selected domains, but they cannot defeat competitive pressure indefinitely. The article assumes that institutions can coordinate against firms and countries competing to automate faster—a direct violation of the Coordination Impossibility premise. It also assumes that developing countries can climb into higher-value work while automation attacks the low-cost service labor that formed their comparative advantage.

The article recognizes externalized electricity and infrastructure costs. It fails to recognize that the deepest externality cannot be priced back onto AI firms: the destruction of mass productive participation.

HIDDEN ASSUMPTIONS

• Productivity gains can be broadly redistributed without the majority losing productive relevance.
• Retraining will produce enough new, scarce roles to absorb workers displaced from routine service and administrative work.
• Employers will accept enforceable labor obligations despite global cost competition and automation arbitrage.
• The Philippines can move from repetitive transactions to analytics, cybersecurity and engineering support faster than AI erodes the lower rungs of that ladder.
• Regulators can act faster and more coherently than capital can deploy, relocate or route around them.
• Public institutions can preserve “human capability” through policy even when firms are rewarded for replacing lower-cost labor.
• Higher-value jobs will expand in sufficient volume rather than becoming narrow positions controlled by a small technical and ownership class.
• International cooperation can overcome the concentration of chips, compute, capital and intellectual property.
• Transfers, retraining and public benefits would preserve more than consumption; in reality, they may simply subsidize a population severed from productive participation.

SOCIAL FUNCTION

Classification: partial truth, transition management and ideological anesthetic.

The article is not pure propaganda. Its warnings about power systems, public cost shifting, labor-market erosion and global inequality are real. But it packages those warnings into a reassuring administrative checklist: make data centers pay, retrain workers, impose labor standards and share the gains.

That script preserves institutional legitimacy. It tells workers they are not being discarded; they are merely “in transition.” It tells governments that enforcement can substitute for ownership. It tells developing countries that moving up the value chain remains available even as automation attacks the very labor arbitrage on which their position was built.

The sedation lies in confusing a slower collapse with averted collapse.

THE VERDICT

Accurate peripheral diagnosis, false central prognosis. The text correctly identifies AI as a global reallocation of electricity, capital, labor and power, and correctly exposes the danger of privatized gains with socialized infrastructure costs. But it stops before the terminal implication: under P1, P2 and P3, better regulation can redistribute some surplus and delay visible breakdown; it cannot restore the mass employment-to-wage-to-consumption circuit once human labor is no longer economically necessary.

The United States may create elite technical and infrastructure roles. That will not absorb displaced Filipino BPO workers—or the broader global labor surplus—at scale. Retraining is a funnel, not a bridge. This editorial audits the scaffolding while refusing to name that the building’s employment function is being removed.