You’ve seen the headlines. You’ve felt the pit in your stomach watching ChatGPT write code, draft legal briefs, and generate art in seconds. We are terrified that artificial intelligence will wake up one day and decide humans are obsolete. But AI doesn’t hate you. AI doesn’t even know you exist. The real threat to your paycheck isn’t a rogue algorithm—it’s a dusty, decades-old tax code.
We spend countless hours debating AI ethics, safety regulations, and alignment. We demand that tech billionaires sign pledges to pause development. But while we’re arguing over science fiction scenarios, the IRS is quietly picking winners and losers. And right now, the tax system is aggressively, structurally hostile to human workers.
The algorithm doesn’t decide who gets fired; the depreciation schedule does.
Think about it from a CFO’s perspective. When a company hires you, they don’t just pay your salary. They pay payroll taxes, healthcare benefits, and unemployment insurance. You are an expensive, high-maintenance asset that can’t be written off. But when that same company invests in an AI system? They get R&D credits. They get to depreciate the software. They benefit from lower capital gains tax rates. The government is literally subsidizing the replacement of human labor with machine capital.
This isn’t an accident or a conspiracy; it’s a system designed for a pre-AI industrial economy. We built a tax structure assuming that capital investment meant building factories that required human operators. Today, capital investment means buying GPUs that replace the operators entirely.
We built an economy that taxes flesh and bone at 30%, and subsidizes silicon at 15%.
This dynamic creates a terrifying paradox. We cheer for AI-driven productivity gains, hoping they will usher in an era of unprecedented prosperity. But our tax system is fundamentally incapable of capturing the value created by these gains. When a human does a job, the government takes a cut through income tax. When an AI does the job, the profit flows up to capital holders, taxed at lower rates, often shielded by pass-through entities. The tax base erodes, public services crumble, and inequality skyrockets.
If you work in tech, policy, or frankly, if you just have a job, you need to understand that the future of work isn’t being decided in Silicon Valley labs. It’s being decided by how we treat depreciation and R&D credits. The specific AI applications that get funded aren’t the ones that are most beneficial to society—they are the ones that offer the best tax shelter.
Stop begging regulators to pause AI. It won’t work, and it misses the point entirely. The only way to ensure a future where humans still have a place in the economy is to rewrite the rules of the game. You can’t regulate a technology into caring about people, but you can absolutely tax a corporation into hiring them.
FAQ
Q: Doesn't AI also create new jobs and new tax revenue?
A: It does create new jobs, but the tax code still penalizes hiring them. The net revenue loss from displaced middle-class labor heavily outweighs the capital gains captured from AI productivity, eroding the broader tax base.
Q: What's the practical implication for businesses right now?
A: If you're a CFO, the math is already done: replacing a human with an AI agent is a tax-advantaged move. If you're a worker, your skills need to be so valuable that the ROI of keeping you outweighs the tax benefits of automating your role.
Q: Should we just implement a 'robot tax' on AI?
A: A direct robot tax is clunky and hard to define. The real fix is leveling the playing field—equalizing the tax burden between human labor and capital investment, and closing loopholes that reward companies solely for shedding headcount.