How your AI agent may level the playing field against companies - The Business Times

TEXT START: As AI agents negotiate on consumers’ behalf, replacing work inside firms, the contest of the next decade is over who keeps the savings

THE DISSECTION

The article identifies two manifestations of agentification: personal agents arbitraging consumer inattention, and enterprise agents substituting for employees. It correctly records the first-order mechanism—agents complete work faster, firms reduce headcount, and the savings become contested.

But it converts a systemic rupture into a pricing question: will competition pass the savings to households or leave them with shareholders? That framing contains the real consequence. The article sees labor substitution, then retreats into the familiar mythology that new jobs, productivity, or lower prices will stabilize the system.

THE CORE FALLACY

It confuses cheaper consumption with economic security. Lower prices are not wages, ownership, or productive participation. A household that saves money through an agent has not gained control of the productive machinery; it has rented an optimization layer from a platform that can be blocked, degraded, price-discriminated against, or monetized.

The two revolutions are not truly pulling in opposite directions. They converge on the same result: capital-controlled agents weaken both the worker’s wage position and the consumer’s bargaining position. Personal agents may temporarily claw back margins, but they do not create a sovereign consumer class. Enterprise agents directly accelerate the severing of labor from income.

The article also treats governance, maintenance, workflow redesign, and human training as possible employment refuges. Under P1, these are not permanent human moats. They are additional functions to be compressed, standardized, verified, and eventually automated. The work may persist temporarily; the labor requirement does not.

HIDDEN ASSUMPTIONS

• Historical technological transitions will repeat, despite AI targeting cognitive work across sectors rather than merely augmenting selected physical processes.
• New jobs will appear faster than existing jobs disappear, and those new jobs will not themselves be automated.
• Consumer agents will remain independent, affordable, and permitted to challenge the firms whose margins they attack.
• Competition will survive concentration in models, data, compute, platforms, energy, and distribution.
• Firms will pass productivity gains to consumers instead of retaining them as rents or using them to fund further consolidation.
• Lower living costs can compensate for the loss of wage income. The article provides no mechanism for that substitution.
• Human oversight and institutional adaptation will remain labor-intensive enough to absorb the displaced majority.
• Deployment delays are treated as opportunities for adjustment rather than as lag before a larger substitution wave.

SOCIAL FUNCTION

Classification: partial truth, transition management, and ideological anesthetic, with an element of elite self-exoneration.

This is not pure copium. It openly admits layoffs, enterprise substitution, and the possibility that AI-generated jobs will also be automated. That makes its diagnosis sharper than ordinary techno-optimism.

Its anesthetic function appears in the conclusion. By presenting the outcome as a contest over who receives savings, it turns a question of ownership and mass dispossession into a question of market competition. Households are invited to hope for cheaper transactions while their role as wage earners is being erased. The article lets the system discuss the distribution of the gains without confronting the death of the employment-consumption circuit.

THE VERDICT

The article detects the weapon but misidentifies the battlefield. Personal agents may create a brief verification-arbitrage window in which attentive consumers recover money from firms. That is a temporary niche, not a reversal of structural power. Firms can close it through access controls, price discrimination, degraded service, or platform ownership.

Under P1, agents achieve durable superiority across cognitive work. Under P2, institutions cannot preserve stable human-only economic domains at scale. Under P3, the majority lose access to economically necessary labor. The post-WWII system therefore dies regardless of whether the immediate spoils go to consumers or shareholders.

The article is ultimately measuring who gets the meat after the animal is already dead.