Alternatively, in our most optimistic scenario, in which AI reduces mortality while also lowering the costs and demand for healthcare, annual deficits fall by 0.8% of GDP in 2044, or about $500 billion per year.

If AI technology leads to a larger old-age population without substantially improving healthcare prices and efficiency, it could increase annual deficits. Our most pessimistic scenario puts 2044 deficits around 1.6% of GDP larger relative to the baseline—about $1 trillion.

And of course, longer lifespans mean more people drawing from SS and Medicare. For example, if AI reduces mortality by 3% (as opposed to 0.74% in the baseline), the old-age population will rise from 73.2 million to 82.4 million in 2044.

A key takeaway is that this technological shock could be different from the internet boom—which boosted productivity and incomes—because AI could also lead to much longer lifespans.

Given the uncertainty around AI, we simulate four scenarios with varying effects of AI on mortality, population, the price of health care, and health care utilization.

Also, check out this Q&A which touches on the main results of our paper: