If AI takes the jobs, who’s left to buy? What stops the collapse and who must act?

Confidence summary: The council is unanimous that government intervention is necessary, but deeply divided on whether to preserve employment or move beyond it entirely.

1. The core argument

When machines can think like lawyers and write like journalists, the feedback loop that sustains consumer capitalism breaks down. China's factories already prove the point: 50 million fewer workers producing more goods than ever before. Now artificial intelligence threatens the white-collar jobs that drive American consumption, which alone represents over two-thirds of U.S. economic output. The mathematics are brutal. If AI eliminates middle-class employment while productivity soars, who purchases the abundance? Markets cannot self-correct when the mechanism of correction itself collapses. Government must step in not as regulator but as economic architect, creating purchasing power through channels that bypass traditional employment. The only question is whether to preserve the fiction of jobs or abandon it for direct distribution of prosperity.

2. How each member frames it

John Maynard Keynes diagnoses this as genuine uncertainty, not calculable risk, making it impossible for markets to price correctly. He draws directly from the liquidity trap analysis that underpinned his General Theory: when private investment disappears because future returns become unpredictable, only government spending can restart the economic engine. But unlike the 1930s depression, AI creates a permanent productivity shift that breaks the traditional link between work and income.

Franklin D. Roosevelt sees the political imperative that Keynes misses. March 1933 taught him that visible action matters as much as economic theory; people need to see government acting decisively when private markets fail them. His New Deal framework applies directly: create public employment that serves genuine social needs while preserving the dignity of work. Social Security succeeded because beneficiaries felt they earned their benefits through contributions.

Amartya Sen fundamentally rejects the premise that employment defines human worth or economic success. Drawing from his entitlements approach developed during Bengal famine studies, he argues the real question is whether people can access capabilities, not whether they hold jobs. AI displacement becomes catastrophic only if we confuse the means of distribution with its ends. The productivity gains exist; the challenge is ensuring everyone shares them.

Albert Hirschman applies his hiding hand principle: we systematically underestimate both the difficulty of transition and our capacity to adapt creatively. His development work in Latin America showed him that rigid planning fails because planners cannot anticipate the institutional innovations that crises force. AI displacement will generate new forms of work and ownership that current policy frameworks cannot imagine.

Milton Friedman accepts that government must act but insists on preserving price signals through direct transfers rather than job creation. His negative income tax proposal from the 1970s becomes more relevant when technological unemployment threatens permanent displacement. Markets will generate new employment categories if government provides income support without distorting labor markets through artificial job protection.

3. Where the council agrees

Government intervention becomes mandatory when private markets cannot maintain purchasing power during rapid technological displacement. All five members accept this despite their different philosophical starting points. They converge on rejecting pure laissez-faire because the current crisis differs qualitatively from previous automation waves: AI targets cognitive work that drives consumption, not just physical labor. The council also agrees that speed matters more than perfection; gradual adjustment risks demand collapse while comprehensive planning awaits perfect solutions. Finally, they recognize that preserving aggregate demand requires either direct income support or public employment, since voluntary private charity cannot match the scale of displacement when artificial intelligence eliminates entire professional categories simultaneously.

4. Where the council splits

The fundamental divide runs between preserving employment relationships and replacing them entirely. Roosevelt and Keynes insist that work provides social meaning beyond income; people need purpose and community, not just purchasing power. They advocate public employment programs that create genuine value through infrastructure, care work, and community services. Against them, Sen and Friedman argue that clinging to obsolete job categories prevents society from realizing AI's full potential. Better to distribute productivity gains directly and let people choose how to spend their time. Hirschman straddles the divide, believing institutional adaptation will eventually reconcile work and income through mechanisms we cannot yet anticipate. Neither side dismisses the other's concerns about human dignity or economic efficiency.

5. For a policymaker to decide on

The choice between universal basic employment and universal basic income cannot be determined by economic analysis alone because it depends on whether society values work as intrinsically meaningful or merely as a distribution mechanism. If work defines human dignity, then government must create jobs even when market demand disappears. If work serves only to allocate resources, then direct transfers preserve freedom while eliminating inefficiency. The policymaker must decide which vision of human flourishing to embed in the institutional response to AI displacement.