AI’s disruption is a choice, not a forecast
When Palantir CEO Alex Karp predicted that AI would erode the economic power of “humanities-trained, largely Democratic voters” in favor of “working class, often male voters,” he wasn’t making a forecast. He was making a choice — and calling it destiny.
That distinction matters more than whether he’s right.
AI has advanced faster than almost anyone expected. Recent geopolitical shocks have compounded the uncertainty. But the real question isn’t who wins or loses in Karp’s vision — it’s whether that vision is the one we want to build.
Disruption Isn’t the Same as Progress
The AI era has generated extraordinary wealth. Nvidia and Microsoft are each worth trillions. ChatGPT now reaches 900 million weekly users. By conventional measures, the revolution is working.
But U.S. unemployment hit a four-year high last November. The wealth gap between the top 1% and bottom 50% has widened since ChatGPT launched. Rapid advancement and record market performance are not measures of success — they’re measures of speed.
A technology capable of unprecedented scientific discovery and work automation should do more than reshuffle economic winners. It hasn’t, largely because industry and government have failed to define what outcomes they actually want AI to deliver — or who it should serve.
Trust Is the Missing Ingredient
People adopted the smartphone because they could see how it would improve their lives. Nobody adopts a technology framed as replacing them.
Yet that’s exactly how some of AI’s loudest advocates describe it. The result is predictable: wariness, skepticism, and a widening gap between immense capability and actual value.
For AI to endure — commercially and socially — people need to trust it. That requires them to feel its benefits directly.
Where AI Should Actually Go to Work
If AI is eliminating manual labor, the economically and socially prudent move is to direct that capacity toward the sectors most starved of it: healthcare, human services, and infrastructure.
These industries face acute labor shortages and stretched staff. They’re also where automating manual tasks would be most transformative without displacing workers:
- Doctors spending more time diagnosing and treating patients instead of documenting visits
- Caseworkers staying in their roles because their jobs no longer consume their weekends
- Transit systems running more reliably as maintenance and reporting become automated
That’s not disruption. That’s progress.
Build With Workers, Not For Them
The US leads in AI talent, research, and infrastructure. The challenge isn’t building the technology — it’s pointing it at the right problems.
One meaningful shift since ChatGPT’s launch: the skills threshold to harness AI has dropped dramatically. LLMs, vibe-coding, and accessible tools mean that building and tailoring technology is no longer reserved for elite college graduates. Frontline workers — the ones who actually understand what’s broken in healthcare or social services — are no longer locked out.
A software engineer knows nothing about being a doctor or a caseworker. If AI is going to serve our most critical workers, industry must build it with them, not for them. Government procurement must do the same. That’s how you get both value and trust.
Stop Predicting. Start Deciding.
Karp is right that AI will reshape economic power. Where he’s wrong is treating that reshaping as inevitable rather than engineered.
The problems most in need of solving aren’t hidden. We know where inequality lives. We know which services are buckling. If the leaders building this technology want it to last, they should stop predicting who gets left behind — and start deciding who gets lifted up.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
This story was originally featured on Fortune.com