Yale's Budget Lab is recommending that policymakers fix the tax code before they start taxing artificial intelligence. The lab argues that addressing existing disparities in the tax code is crucial for equitable AI-driven economic growth and balanced fiscal benefits.
Why the tax code comes first
The recommendation rests on a simple premise: the current tax code has built-in disparities that could distort how the gains from AI are distributed. If those disparities are left untouched, the lab says, any new taxes on AI would likely hit some groups harder than others, widening the gap between those who benefit from automation and those who don't.
By fixing the code first, the lab argues, policymakers can create a level playing field. That means closing loopholes, adjusting rates, and making sure the tax base reflects the modern economy before layering on AI-specific levies. The goal is to ensure that the fiscal benefits of AI growth are balanced across the board, not concentrated in a few sectors or income brackets.
What the recommendation means
This is a recommendation, not a policy. It doesn't spell out which taxes to impose on AI or how to structure them. Instead, it sets a sequence: repair the foundation before building on top of it. The lab's point is that a fair tax code is a prerequisite for any AI tax to work as intended.
That sequencing matters. If AI taxes are introduced while the code still has disparities, the lab warns, the result could be uneven fiscal outcomes. Some companies might find ways to shift their AI-related income into lower-taxed categories, while others—especially smaller firms or workers—could end up carrying a heavier load. Fixing the code first reduces that risk.
The debate over AI taxes
The recommendation lands in the middle of a growing conversation about how to tax artificial intelligence. Some proposals have called for direct taxes on AI systems or on the productivity gains they generate. The lab's contribution is to argue that the existing tax code needs repair before any of those ideas can be implemented fairly.
That's a different angle from the usual debate. Instead of asking whether to tax AI, the lab is asking what the tax code looks like underneath. It's a reminder that new taxes don't exist in a vacuum—they interact with the rules already on the books.
The recommendation doesn't set a timeline, but it gives policymakers a clear order of operations. Any move to tax AI will likely have to address the tax code first, and the lab's argument is that doing so is the only way to ensure the benefits of AI are shared fairly.
