The rapid evolution of artificial intelligence has sparked a global debate, and tax policy is the latest arena for discussion. As fears mount regarding AI’s potential to disrupt the workforce and reshape the economy, several high-profile figures—including U.S. Senators and industry leaders—have suggested that the tax code should be used to either slow AI adoption or mitigate its social side effects.
However, crafting tax policy based on extreme, speculative scenarios is a strategic error. Instead, the smartest path forward is to implement robust tax reforms that provide value regardless of how the AI revolution unfolds.
The Uncertainty of the AI Impact
Current predictions about AI range from modest productivity gains to the total replacement of human labor. While theoretical models of a “jobless future” are interesting for academic study, they aren’t a reliable foundation for immediate legislation.
To date, there is no empirical evidence of a massive, AI-driven labor displacement. While the entry-level white-collar market has seen a slowdown, current labor statistics remain historically strong, with prime-age labor force participation reaching levels unseen for much of the 2010s. The current shift looks less like a total replacement and more like the arrival of the personal computer—a transformation, not an extinction.
Why Targeted AI “Excise Taxes” Are a Bad Idea
Several proposals suggest creating “AI Horizon Funds” or imposing excise taxes on “token generation” and automation. The goal is often to protect human workers by making technology more expensive. This approach is flawed for several reasons:
-
Automation is Not “Privileged”: Contrary to some arguments, the tax code does not unfairly favor machines. Businesses can fully deduct human wages immediately. However, they are often forced to spread out deductions for hardware and software over many years, which actually penalizes capital investment due to inflation and the time value of money.
-
Stifling Growth: History shows that taxing technological inputs—like the railroads in the 19th century or software in the 1990s—leads to economic stagnation. Recent data from the Federal Reserve suggests that the growing GDP gap between the U.S. and Europe is partially due to slower tech investment in Europe.
-
The Global Race: Nations that hesitate to adopt new technologies generally fall behind. Designing policy specifically to slow down innovation is a recipe for long-term economic decline.
Better Alternatives: Reskilling and UI Reform
Instead of punishing innovation, policymakers should focus on making the labor market more flexible and supporting workers through transitions.
-
Unemployment Insurance (UI) Reform: The current U.S. “experience rating” system can discourage companies from hiring “risky” workers, such as those switching careers. Moving away from this model could make it easier for displaced workers to find new roles in emerging sectors.
-
Boosting Worker Training: Current tax laws cap the amount companies can deduct for investing in employee education. Raising these caps and streamlining workforce development credits would encourage businesses to retrain their existing staff for an AI-integrated workplace.
AI and the Deficit: Two Perspectives
There are two competing narratives regarding how AI will affect national revenue:
-
The Pessimistic View: Automation will shrink the pool of taxable human income, leading to a revenue collapse that necessitates radical new taxes on AI companies.
-
The Optimistic View: AI will spark a productivity boom, increasing overall incomes and corporate profits, thereby naturally increasing tax revenue and shrinking the deficit.
History leans toward the optimistic view. Since the 1940s, labor’s share of net income has remained remarkably stable at around 70%. If AI boosts productivity, total income rises, and tax receipts follow suit. Even if AI only helps the economy meet current baseline growth projections, it stabilizes the fiscal outlook rather than destroying it.
A Move Toward Consumption-Based Taxes
While we shouldn’t reinvent the wheel for AI, the shift toward AI-driven profits makes traditional tax reforms even more attractive.
The U.S. remains the only developed nation without a Value-Added Tax (VAT). Shifting toward consumption-based taxation provides a more stable revenue stream than relying solely on income. Furthermore, ensuring that the Corporate Income Tax allows for the full and immediate deduction of all capital and labor costs would ensure we are taxing true “supernormal” profits without discouraging the investment that drives progress.
Conclusion
AI may eventually transform society, but until “science fiction” scenarios become economic reality, we should stick to proven principles. Tax policy should encourage growth, support worker mobility, and rely on stable broad-based consumption—rather than trying to predict the unpredictable.