Beyond the AI Wealth Pitch: Real Worker Protections Now

Ramaswamy says workers can build wealth in the AI era. He makes a hopeful, market-friendly case that new tools will open routes to ownership, higher-paying complements to automation, and more entrepreneurial upside for people outside the executive suite. I’ll grant the optimism where it’s warranted — but he skips the part where the rules of the game determine who actually cashes the checks.

He’s right about one thing: ownership matters. Wealth tends to accrue to those who own assets, not just those who rent out their time. Stock, profit-sharing, equity in startups — these are how people ride big technology waves. But that’s only half the story. Getting a stake and turning that stake into real, liquid value are two very different problems, and Ramaswamy mostly treats them as the same one.

Start with who we’re talking about.

“Workers” are not a single category you can train into prosperity. The bargaining position of a software engineer at a fast-growing tech firm is not the bargaining position of a call-center rep, a warehouse picker, or a home health aide. Telling all of them to “partner with AI” collapses wildly different starting points into one feel-good slogan.

Equity and profit-sharing sound democratizing on paper. They work — sometimes — in companies that already offer them and where workers have enough bargaining power to negotiate terms that aren’t symbolic. For gig workers juggling unstable hours, or hourly staff boxed in by franchise models and noncompetes, that story doesn’t translate. Let’s be real: if your job can be scheduled by an app, your odds of being handed meaningful ownership are close to zero unless someone changes the structure of the deal.

From my years at Goldman, I learned to ask two questions whenever someone pitched a “shared upside” story: who sets the terms, and who can walk away? The people with capital, legal teams, and networks usually design the upside so they stay at the top of the waterfall. Telling everyone else to just “skill up” without touching capital access is like telling someone to run a marathon while you quietly remove their shoes.

Then there’s control of the AI stack itself. The systems are being built, trained, and distributed by a relatively small set of platforms and large incumbents. They own the models, the data pipelines, and the channels. That concentration shapes who can build viable companies on top. Workers may get stock options at those firms, but whether those options mean wealth or wallpaper depends on decisions they don’t control: growth-at-all-costs strategies, repeated fundraising, and exit timing that can turn “ownership” into a long-dated lottery ticket.

The same dynamics show up in geography and education. It’s easier to translate AI “opportunity” into actual equity if you sit in a major tech hub, know people who’ve already done it, and can live without a paycheck for a while. Training programs that look visionary in op-eds collide with basic realities: long commutes, childcare gaps, no spare savings, spotty broadband. You can’t reskill on an empty bank account.

That’s why policy and corporate design quietly sit underneath every cheerful story about AI and worker wealth. If the private sector is the engine, policy is the steering wheel: tax rules that decide how stock compensation gets treated, antitrust enforcement that either tolerates or pushes back on concentration, employment law that defines who even counts as an “employee” eligible for benefits and equity. Ramaswamy wants markets to deliver; markets can, but only if the rulebook doesn’t structurally reward scale over diffusion.

He also leans hard on a clean line from skills to bargaining power. Train people to work alongside AI, he argues, and they’ll capture more value. Sometimes. But not every AI-related skill is scarce, and not every employer will pay a premium just because someone can prompt a model or sit next to a chatbot. If firms can swap in cheaper capital or offshore labor, the individual “upskilled” worker still has to compete against that option.

What really shifts bargaining power is a mix of skills and alternatives: the ability to switch jobs, move locations, tap financing, or negotiate collectively. Without those, human capital is just a better-decorated resume. Expanding access to capital — via employee ownership models, real liquidity for worker-held shares, or cooperative structures — isn’t some automatic byproduct of AI adoption. It’s a political and corporate choice.

There’s also a historical pattern here that his argument glides past. When mechanization reshaped manufacturing, or when software rolled through back offices, the gains didn’t automatically flow to workers — they flowed to firms and shareholders until unions, regulation, and new norms forced a different split. AI is shaping up the same way: huge productivity upside, but absent countervailing pressure, that upside defaults to the people and companies already holding the equity.

Look at how some firms are already drawing the map. A tech giant can deploy AI to thin out middle-management and support roles, then offer remaining staff “AI training” and a sprinkle of stock-based comp. On paper, everyone is more “productive.” In reality, the company saves on labor, shareholders see the bump, and a smaller core workforce takes on more responsibility with only marginally better odds of building lasting wealth.

Critics of this view will say that’s too cynical. They’ll point to AI-fueled startups, new job categories, and the long track record of technology waves creating more opportunity than they destroy. Right, entrepreneurs do create value. Some workers will absolutely ride AI into better roles, solid equity, and real upside. But that doesn’t answer who’s left out or how much friction stands between a theoretical path and a practical one: regulatory hurdles, platform lock-in, and network effects that tilt new markets toward incumbents before small players can breathe.

You can absolutely celebrate innovation and still argue that ownership needs to be designed, not assumed. Training matters. So do stock plans that aren’t illusory, legal definitions that don’t shove workers into the gray zone, and capital channels that don’t require already knowing the right people.

Ramaswamy is right that AI can expand the pie for workers who get meaningful stakes in the value they help create. The more interesting question — and the one his column only brushes past — is who writes the rules that decide how those slices get cut.