The Economic Impact of AGI

Economic theory suggests that a technology capable of performing all cognitive labor would, in the limit, produce growth rates far above historical norms. Standard models break down at this point because they assume labor as a scarce input.

Empirically, current AI tools already produce measurable productivity gains — 14% on average for customer support agents (Brynjolfsson et al., 2023), 56% for software engineers using AI assistants (GitHub, 2024).

Goldman Sachs estimated in 2023 that generative AI could expose roughly 300 million full-time jobs to automation globally, with two-thirds of US occupations seeing some degree of automation.

Historical technology transitions (electrification, computing) ultimately produced net job creation, but transition periods imposed significant costs on displaced workers. The pace of AI deployment is significantly faster than past transitions.

The distributional question — who captures the gains — may matter more than the aggregate question. Productivity gains can flow to capital owners, to skilled workers complementary to AI, or to consumers via lower prices.

Without intentional policy, gains tend to concentrate. Universal basic income, capital broadening, and revised tax structures are under active discussion in policy circles.

If AGI eventually performs all economically valuable labor, the central economic question shifts from production to distribution and meaning. Whether this transition is utopian or dystopian depends almost entirely on how political institutions adapt.