Does the AI Revolution Help or Hurt People?

Originally published on Substack.

If you work in policy, government, or research, this moment should feel familiar.

A new technology emerges. Headlines turn anxious. Hearings follow. White papers multiply.

And before the technology has fully diffused—or its benefits broadly reached workers and consumers—policymakers begin debating how to slow it down, fence it in, or “protect” people from it.

That’s exactly where we are with artificial intelligence.

And the biggest risk right now isn’t AI moving too fast. It’s policy reacting too quickly—and getting the incentives wrong.

The Economist acknowledges that AI is not the only—or even the main—threat facing large technology firms. Regulation, antitrust pressure, and trade policy loom larger. That should give policymakers pause. When the most powerful firms in the economy fear government more than competitors, something fundamental has gone wrong.

The same analytical error shows up in the labor market.

Concerns that AI will eliminate entry-level jobs are understandable. But as The Economist explains in its assessment of AI’s impact on entry-level work, AI tends to reshape tasks rather than eliminate career ladders—just as past general-purpose technologies did. The difference between disruption and opportunity lies in whether labor markets are flexible enough to adjust.

Here is where policy matters most.

Entry-level workers are often blocked not by automation, but by occupational licensing, credential inflation, rigid labor rules, and wage mandates that raise the cost of hiring inexperienced workers. When legal risk and fixed costs rise, employers substitute capital for labor more quickly. Technology then gets blamed for outcomes regulation helped create. That is not a market failure—it is a policy-induced distortion.

Misplaced blame extends to AI leadership as well.

Another Economist article pushes back on the caricature that AI executives are uniquely dangerous, noting that AI bosses are just regular capitalists—responding to incentives, risk, and expected returns. Capital is not ideological. It flows where rules allow it to flow. When policy rewards scale, compliance, and political access, capital concentrates. When barriers to entry fall, competition expands.

This distinction matters for anyone designing regulation.

Too often, we confuse free-market capitalism with policy-engineered concentration. The former is disruptive and decentralized. The latter is orderly, cartel-friendly, and politically tempting. If policymakers want less concentration, the answer is not layering regulation onto incumbents—it is removing rules that block challengers.

Banking provides a clear example.

As the Wall Street Journal reports on bank branches closing across the country, the trend is often framed as abandonment of communities. But the real drivers are technological change combined with regulatory cost.

Digital banking lowers operating expenses and expands access. Compliance requirements raise fixed costs—costs large institutions can absorb and community banks often cannot. Consolidation follows. Then markets are blamed for outcomes policy accelerated.

Trade policy reinforces the same mistake.

Despite mounting evidence of harm, The Economist asks whether the U.S. has reached “peak tariff”. Tariffs persist because they are politically attractive, even though they function as hidden taxes on consumers and producers. For AI, advanced manufacturing, and modern supply chains, tariffs raise input costs, slow diffusion, and weaken resilience—yet remain a favored tool.

Put together, the lesson is clear:

AI is not the force concentrating power in the economy. Policy is. When entry is restricted, incumbents win. When compliance is expensive, scale is rewarded. When prices are controlled, supply contracts. When subsidies expand, inefficiency hides.

Markets discipline failure quickly. Politics rarely does.

Closing Thoughts

For policymakers and analysts, the choice is straightforward.

We can treat AI as a threat to be contained—or as a tool whose benefits depend on competition, openness, and institutional humility.

Basic economics offers a clear guide: people respond to incentives, and markets aggregate information better than governments ever can.
​
AI will advance regardless. The only open question is whether policy will amplify its benefits—or suffocate them before they spread. Thanks for reading.