Anthropic co-founder says AI could boost US GDP by 50% by 2030 - Daily Jang

TEXT START: Anthropic co-founder Jack Clark has issued a warning to policymakers that rapidly advancing artificial intelligence (AI) could transform the global economy, with one of the company’s internal models estimating that US economic output could rise by nearly 50% by 2030 under specific conditions.

The Dissection

The article packages an extreme growth forecast as both promise and warning. It treats AI-driven scientific discovery, infrastructure, regulation, job losses, and catastrophic misuse as parallel issues without confronting their central contradiction: AI may expand output while eliminating the wage-based participation that stabilizes the existing system.

The estimate comes from an unnamed internal model and unspecified conditions. The article supplies no methodology, ownership assumptions, distribution mechanism, or explanation of how new industries would absorb displaced workers. It presents GDP expansion as the headline reality and employment collapse as an unresolved footnote.

The Core Fallacy

The article confuses more output with system survival. GDP measures aggregate production; it does not measure who owns the AI capital, who receives the resulting income, or whether humans remain economically necessary.

A 50% GDP increase could intensify the Discontinuity Thesis. AI can create new industries while automating those industries as well. The claim that new inventions would be required for such growth does not imply that those inventions would require mass human employment. Under P1, cognitive automation raises productive capacity. Under P3, it can simultaneously destroy productive participation. Growth is not exoneration.

Hidden Assumptions

• The gains from AI will be broadly distributed rather than captured by AI-capital owners.
• New sectors will create enough human employment to replace the work destroyed by automation.
• Wage income can remain the primary distribution mechanism despite falling human necessity.
• Regulation and international cooperation can keep pace with AI capability and prevent coordination failure.
• Data centres, energy, logistics, maintenance, and physical infrastructure will scale without becoming binding constraints.
• Human oversight will remain competent as systems become more capable.
• GDP growth will translate into social stability rather than a larger surplus controlled by a narrower class.
• The phrase specific conditions conceals assumptions that have not been tested or disclosed.

Social Function

Classification: partial truth, elite self-exoneration, and transition management.

The article acknowledges job losses and catastrophic misuse, which gives it a sober surface. Its deeper function is to make displacement politically tolerable by promising abundance. It shifts attention from ownership and control to national output, allowing policymakers and capital owners to describe mass economic exclusion as the price of progress.

The regulatory appeal is not worthless, but it does not answer the structural problem. Safeguards may reduce misuse; they do not restore the mass employment-to-wage-to-consumption circuit once AI makes human cognitive labor dispensable.

The Verdict

This is not evidence that AI will save post-WWII capitalism. It is evidence that AI may enlarge the productive surplus while severing the mechanism that distributes purchasing power through employment. If the forecast materializes, the central question will not be whether the economy grows. It will be who controls the machines, who receives the output, and whether ordinary humans remain economically necessary. Without that answer, the 50% figure is not a rescue plan. It is the projected size of the carcass.