AI’s $2.2 trillion deficit fix is already half fake, economists say

AI could shave $2.2 trillion off the U.S. deficit by 2036. But according to a new working paper from economists at Brookings and the Federal Reserve, more than half of that savings could vanish — canceled out by the very disruption AI itself would cause. In May, the U.S. national debt crossed the eye-popping $39 trillion mark. The difference between what the U.S. government spends and what it earns has become a galvanizing force for fiscal hawks of all political stripes. Without significant reform measures from Congress, the widening deficit threatens to deplete the trust funds that finance Social Security in 2032, and Medicare one year later. Budget experts say fixing the deficit will require tax hikes, cuts to entitlement, or most likely, a combination of both. Absent that political will, AI has been floated as a fiscal escape hatch. A new paper suggests that escape hatch is narrower than advertised. If AI leads to a major increase in productivity, higher output per worker across the economy could help boost government coffers and stabilize the budget, according to a working paper published Wednesday by the Brookings and Fed economists. On the revenue side, AI-driven productivity gains would mean the government can collect more from a bigger economy without necessarily raising tax rates. On the spending side, AI could also help erase inefficiencies, particularly in health programs, where administrative costs alone account for one quarter of all expenses. In total, an AI productivity surge could reduce the country’s annual budget deficit from the roughly 6% of GDP it currently sits at to as low as 2%, the equivalent of $2.2 trillion wiped clean from America’s bill by 2036, the paper’s authors wrote. But that number comes with an immediate caveat the paper’s authors bury in their conclusion: the same AI boom could claw back more than half of those savings through five compounding side effects. Technology has delivered such miracles before. In the 1990s, the In