Artificial intelligence could deliver the productivity surge policymakers have been hoping for since the global financial crisis. But even if it does, economists caution that faster growth will not be enough to solve the mounting debt burdens weighing on advanced economies.
Public debt already exceeds 100% of GDP across most rich nations and is projected to rise further as ageing populations strain pension and healthcare systems, interest bills climb and governments ramp up defence and climate spending. Against that backdrop, AI is increasingly being framed as a potential fiscal lifeline.
The reality is more complicated.
Productivity: The “Magic” Ingredient-With Limits
Economists broadly agree that sustained productivity growth can dramatically improve fiscal dynamics. Higher output boosts tax revenues without raising tax rates, makes existing debt easier to service and reassures bond investors worried about long-term solvency.
At the Organisation for Economic Co-operation and Development (OECD), modelling suggests that if AI meaningfully raises labour productivity and if employment also expands public debt across member countries could be about 10 percentage points lower by the mid-2030s than otherwise projected. Even then, debt would still climb to roughly 150% of GDP on current trajectories, up from around 110% today.
In the United States, best-case projections from several economists suggest debt could rise more gradually, to roughly 120% of GDP over the next decade rather than accelerating more sharply. But that still represents historically elevated levels.
As one economist put it, productivity is “like magic” for fiscal sustainability yet today’s debt challenges are too large for productivity gains alone to offset.
Demographics: The Structural Headwind
The fundamental constraint is demographic.
Ageing populations mean fewer workers supporting more retirees, pushing up pension and healthcare costs. In the United States, Social Security alone accounts for roughly one-fifth of federal spending, and benefits are indexed to wages. If AI lifts wages, it may simultaneously increase future benefit obligations.
Slowing immigration in some countries, particularly the U.S., compounds the issue by limiting labour force growth. If AI boosts output per worker but the total number of workers stagnates or declines, overall fiscal relief may be limited.
In short, AI may buy time but it does not reverse the demographic arithmetic driving long-term deficits.
Growth vs. Interest Rates: A Delicate Balance
For debt sustainability, what matters is not just growth, but the relationship between growth and borrowing costs.
If AI-driven productivity pushes economic growth above interest rates for a sustained period, governments can stabilise or even reduce debt ratios more easily. But if faster growth also lifts real interest rates for example, because higher productivity raises returns on capital then debt servicing costs could rise in parallel.
This debate is already unfolding among policymakers at the Federal Reserve, where officials are assessing whether AI could permanently raise the economy’s potential growth rate.
Bond markets will be decisive. Since the pandemic, investors have shown a willingness to punish governments perceived as fiscally profligate. Higher yields can quickly offset any growth dividend from technological gains.
Employment and Wages: The Distribution Question
Much depends on how AI reshapes labour markets.
If AI complements workers and creates new categories of employment, tax revenues may rise meaningfully. But if automation displaces workers faster than new jobs are created, or if profits accrue disproportionately to capital rather than labour, fiscal gains could disappoint.
Capital income is often taxed more lightly than wages. A productivity boom concentrated in corporate profits rather than payrolls may widen inequality without generating proportionate public revenue.
On the spending side, governments might benefit from efficiency gains in public administration. Yet history suggests higher growth can also lead to higher spending demands from infrastructure upgrades to social transfers.
Even in optimistic scenarios where AI lifts U.S. growth closer to 3% annually for an extended period, debt ratios are projected to stabilise at elevated levels rather than return to pre-crisis norms.
In pessimistic scenarios where AI disappoints or a recession strikes before productivity gains materialise debt trajectories could worsen significantly, potentially reaching levels that trigger market instability.
The consensus among economists is clear: AI can ease fiscal pressure, but it cannot substitute for structural reforms. Addressing entitlement sustainability, improving tax efficiency and managing spending priorities remain central.
A Race Against Time
There is also a sequencing risk. If financial markets grow nervous about fiscal trajectories before AI-driven gains are realised, borrowing costs could spike. In that case, the productivity dividend may arrive too late to calm bond investors.
Technological revolutions historically take time to diffuse across economies. Infrastructure, regulation, workforce training and corporate adoption all shape how quickly productivity benefits materialise.
For debt-laden economies, the gamble is that AI’s boost will be large, broad-based and timely. That is possible but far from guaranteed.
AI may help governments breathe easier. It will not absolve them of the harder political choices required to put public finances on a sustainable path.
With information from Reuters.
