A New Anthropic Model Seeks to Test How AI Could Impact the US Economy | KQED

TEXT START: Will artificial intelligence light a fire under the U.S. economy in the coming years?

1. The Dissection

The article converts a possible labor-system rupture into an interactive macroeconomic exercise. Its variables—AI capability, adoption speed, worker replacement, and reemployment—make structural destruction look like adjustable model settings.

It admits a severe scenario: GDP surges, nearly 14% of workers lose their jobs, and fewer than half find new work. But it stops at output, employment, and tax revenue. It does not examine ownership, bargaining power, institutional legitimacy, or whether human labor remains economically necessary. “Policymakers should get ready to spend” is the governing move: labor exclusion is reframed as a fiscal-management problem.

2. The Core Fallacy

The article treats higher GDP and tax revenue as if they can repair the loss of productive participation. They cannot. Transfers may preserve consumption; they do not restore wage-based necessity, bargaining power, or control over production.

It also treats reemployment as a meaningful stabilizing variable without confronting the possibility that AI eliminates economically necessary cognitive work faster than new human tasks can emerge. Adoption friction is a lag defense, not a refutation. If the technology becomes decisively cheaper and better, competitive pressure eventually forces diffusion.

The article measures whether the economic engine grows. It does not ask whether humans remain part of the engine.

3. Hidden Assumptions

• Enough new, economically necessary work will appear for displaced workers.
• Governments will successfully tax AI-generated gains and distribute them at sufficient scale.
• AI adoption will remain slow, controllable, and institutionally manageable.
• Productivity gains will diffuse broadly rather than consolidate ownership and power.
• GDP growth will translate into social stability and legitimacy.
• Governments can spend their way through displacement without political conflict or elite resistance.
• AI capability growth will not recursively accelerate adoption and further substitution.
• Public expectations and a survey of nearly 11,000 people provide meaningful evidence about structural economic outcomes. They provide sentiment, not system mechanics.

4. Social Function

Primary classification: ideological anesthetic and transition management, with a secondary function as partial truth and elite self-exoneration.

This is not pure propaganda. The article acknowledges unemployment and an extreme transformation scenario. Its anesthetic function comes from relocating the crisis into scenario parameters, diffusion speed, and tax-funded spending. The refusal to predict protects the institution from being wrong while normalizing the idea that any consequences remain administratively soluble.

The message is: build the replacement system first; let policymakers fund the displaced afterward.

5. The Verdict

This is a useful dashboard built on an inadequate gauge. It correctly identifies capability, adoption, substitution, and reemployment as relevant variables. It fails to confront the terminal question: whether human labor remains economically necessary at scale.

Under the Discontinuity Thesis, if P1, P2, and P3 mature, higher GDP does not save post-WWII capitalism. It finances the transition from a wage society to owner-controlled production with subsidized consumers. Spending can delay social death and suppress immediate unrest. It cannot restore productive participation.

The article’s extreme scenario is therefore not a solution. It is the first clean bureaucratic description of the corpse.