Silicon Valley’s Cash Experiments Test AI’s Human Toll

SANTA CLARA, Calif. — Mikole Weeks got the call in December 2024. A guaranteed income program would send her $1,330 every month. No questions asked. No work requirements. For one year.

She bought softer diapers for her children. Paid for clean laundry. Took them to Six Flags. Simple things. Yet they changed her days. Weeks told The New York Times the money eased constant worry. It let her focus. Her younger sister, inspired, enrolled in college too.

Stories like this surface across Santa Clara County. Ten pilots now run here. They reach more than 980 households. Some target former foster youth. Others support unhoused families or young parents. Payments hover around $1,000 monthly. Google funded six of them in 2022 with $50 million. The goal? Blunt homelessness in one of America’s priciest regions. (Guaranteed Income SV)

But these local trials carry heavier weight now. Artificial intelligence races ahead. It automates code. It tackles mathematics once thought untouchable. Tech leaders who build these systems warn of sweeping job losses. Their answer, increasingly, involves cash. Direct. Unconditional. From government or industry profits.

Elon Musk calls it universal high income. Not mere survival. Abundance. In April he posted on X that federal checks offered the best response to AI-driven unemployment. By September, standing beside President Trump, he painted a bolder picture. “I think by far the most likely outcome is an age of abundance, where we don’t have just universal basic income, we have universal high income.” Jobs would change, he said. They always have. Robots and superintelligence could deliver better medical care than anyone enjoys today. Wonderful, he added.

Sam Altman once poured millions into UBI research. His OpenResearch project tracked 3,000 participants. Early findings showed gains in mental health, reduced food insecurity, even slight boosts to short-term employment. A separate study of 1,000 low-income adults receiving $1,000 monthly found participants valued work more, not less. Yet Altman has stepped back. In interviews this year he expressed less faith in fixed cash alone. Collective ownership of AI gains excites him more. Everyone must share the upside. (The New York Times)

Dario Amodei of Anthropic suggested taxing AI companies to fund payments. Andrew Yang, once a presidential candidate on this single issue, still pushes variations. Even California Gov. Gavin Newsom directed officials to examine universal basic capital — giving workers ownership stakes in the technology disrupting their livelihoods.

Local Results Meet Looming Disruption

Reality on the ground looks less utopian. Jennifer Loving runs Destination: Home, a nonprofit overseeing several Bay Area programs. She told The New York Times it’s a lot to ask people to rebuild their lives in 12 months. Many problems remain structural. Systemic. Cash helps. It does not rewrite housing markets or education systems.

The Silicon Valley Guaranteed Income Project, a randomized trial by UCSF’s Benioff Homelessness and Housing Initiative, tracks 300 families. Some receive $1,000 monthly for 24 months. Others get usual services. Primary completion arrives mid-2026. Early signals echo broader research. Participants spend more on family and friends. Employment doesn’t collapse. Mental health ticks up. But scalability? Tax laws cap durations and amounts. Permanent programs would demand new revenue. (ClinicalTrials.gov)

Critics push back hard. Sridhar Vembu, founder of Zoho, rejected the entire concept this week. In a post responding to OpenAI’s Andrej Karpathy, he called pure UBI “unimaginative and dystopian.” It assumes monopolies will control production. Governments then compensate idle citizens. Vembu prefers distributing automation’s gains so people earn through raising children, caring for elders, restoring nature or making music. Those activities should pay well when machines handle nearly all labor. (Rediff.com, India Today)

Some organizers already target AI-displaced workers. A coalition including the Fund for Guaranteed Income and tech worker advocates launched AI Dividend. Nearly 50 participants receive up to $1,000 monthly plus training. Initial funding hit $300,000. The program runs a year. It’s not pure UBI. Conditional. Time-limited. Aimed at people whose skills no longer match the market. Many come from layoffs at companies like Oracle. (Business Insider)

Yet momentum builds for broader ideas. A Jerusalem think tank founder modeled Universal High Income using current AI tools. Tax AI production. Create public wealth funds. Citizens gain stakes in leading tech firms. One estimate floated $60,000 annually to support a middle-class life without traditional employment. The Economic Times explored the concept just yesterday, noting the debate has moved past theory. (The Economic Times)

Stanford researchers updated payroll data in August. Employment for young workers in AI-exposed roles lagged 19 percent behind less-exposed ones. Correlation, not yet causation. Still. The pattern raises alarms. A House bill from late 2025 proposed $250 monthly for qualifying adults. It sits stalled.

But. The pilots reveal human dimensions lost in grand theories. One mother used funds to stabilize housing. Another paid off small debts that once paralyzed her. College enrollment rose in families touched by the payments. These aren’t abstract productivity statistics. They are families navigating uncertainty.

So what happens when AI erodes not just routine work but cognitive roles once considered safe? Software. Analysis. Even advanced math. Vembu calls it humbling. Pure mathematics yields to machines. Many intellectual jobs follow.

Proponents see shared wealth funds as the answer. OpenAI itself published a framework for a public fund financed by tech companies. Citizens gain stakes in AI growth. Newsom’s directive studies similar concepts. Universal basic capital. Ownership, not just checks. (The New York Times)

Opponents warn of dependency. Reduced incentive. Chicago’s recent flirtation with $500 monthly for 1,000 families drew fresh scorn from fiscal conservatives. Yet experiments continue. Santa Clara County’s efforts, backed by local government and philanthropy, test targeted help for justice-involved individuals, low-income seniors, unhoused students.

The data so far paints a cautious picture. Short-term gains dominate. Mental health improves. Food insecurity drops. Employment holds or rises modestly. Long-term effects on labor participation remain unclear. Tax limits force pilots to end. Participants know the clock ticks.

Musk’s abundance vision assumes production costs collapse so dramatically that goods become nearly free. Healthcare for all exceeds today’s standards for billionaires. Money itself might matter less. But he offers no detailed mechanism. How do societies tax or redistribute when robots dominate output and payroll taxes shrink?

Bill Gates has urged taxing robots directly. Others propose sovereign wealth funds owning AI assets on behalf of citizens. Daniel Schreiber’s models explore production taxes funneled to direct transfers.

These debates feel urgent because the technology does not wait. AI models improve monthly. Companies deploy them aggressively. Layoffs in tech already ripple. White-collar roles once immune now face scrutiny.

Local programs in Silicon Valley offer a laboratory. They show cash can stabilize. It buys time. Breathing room for retraining or family focus. Yet they also expose limits. One year passes quickly. Structural barriers remain. And scaling to millions demands political will, new revenue streams and acceptance that work itself may transform.

Altman captured the tension. Cash helps. Participation in upside matters more. Ownership. Stakes. Not passive support. Vembu goes further. Reject the dystopia of compensated idleness. Create conditions where human pursuits retain economic value.

The coming years will test these visions. More pilots will report. AI capabilities will surge. Policymakers in Sacramento and Washington will study options. Universal basic income. Universal high income. Universal basic capital. Hybrids. Each carries assumptions about human nature, economic incentives and political feasibility.

Weeks and her family experienced one narrow slice. Softer diapers. Family outings. Renewed hope. Multiply that by thousands. Then ask whether such relief scales to an economy where machines outperform humans across domains. The answer remains unwritten. But the experiments continue. Cash flows. Data accumulates. And the machines keep learning.