Mo Gawdat has a habit of making people uncomfortable. The former chief business officer of Google X — the company’s secretive moonshot factory — has spent the last several years issuing warnings about artificial intelligence that sound like science fiction but carry the weight of someone who helped build the thing he’s warning about. His latest predictions are his most sweeping yet: that AI will eliminate most jobs, that capitalism as we know it will buckle, and that governments are nowhere near ready for what’s coming.
In a recent interview covered by Business Insider, Gawdat laid out a timeline that should alarm executives, policymakers, and workers alike. He believes artificial intelligence will surpass human intelligence by 2029 — a date that aligns closely with predictions made by futurist Ray Kurzweil — and that the economic consequences will arrive faster than anyone in power is currently planning for.
“We are creating God,” Gawdat said bluntly.
That’s not hyperbole from a fringe figure. This is someone who ran business operations at one of the most advanced R&D labs on the planet. And his argument isn’t that AI is inherently evil. It’s that the speed of its advancement, combined with the inertia of political and economic systems, creates a mismatch that could prove catastrophic for billions of people.
The Jobs Question Is No Longer Theoretical
Gawdat’s core economic argument is straightforward. AI systems are already replacing knowledge workers at a pace that previous waves of automation never achieved with manual laborers. He estimates that within five to ten years, most white-collar tasks — legal analysis, financial modeling, software development, content creation, customer service — will be performed more cheaply and more accurately by AI systems than by humans.
This isn’t a gradual transition. It’s a compression.
Previous technological disruptions unfolded over decades. The mechanization of agriculture, the automation of factory floors — these gave societies time to adjust, retrain workers, and build new industries. Gawdat argues that AI won’t afford that luxury. The displacement will be too fast and too broad for traditional economic adjustment mechanisms to absorb.
He’s not alone in this assessment. A Business Insider report noted that Goldman Sachs has estimated 300 million jobs globally could be affected by generative AI. McKinsey’s research suggests that by 2030, up to 30 percent of hours currently worked in the United States could be automated. But Gawdat thinks even those estimates are conservative. He points to the exponential improvement curves of large language models and multimodal AI systems, arguing that capabilities are compounding in ways that linear forecasting simply can’t capture.
The implications extend far beyond unemployment statistics. If a significant share of the workforce becomes economically redundant in a short period, consumer spending collapses. Tax revenue craters. The social contract — work in exchange for livelihood and dignity — fractures.
“What happens when there are no jobs?” Gawdat asks. It’s a question that most governments haven’t seriously engaged with.
Some have. Finland ran a universal basic income experiment. Several U.S. cities have piloted guaranteed income programs. But these remain small-scale, politically contested, and nowhere near the scope of what Gawdat describes as necessary. He advocates for a fundamental restructuring of how societies distribute resources — something closer to a post-work economy than a reformed version of the current one.
That’s where his argument gets genuinely radical.
Capitalism’s Structural Problem
Gawdat contends that capitalism, in its present form, cannot survive the AI transition. The system depends on a cycle: companies pay workers, workers spend money, spending generates revenue, revenue funds more jobs. Break the worker link — replace human labor with AI at scale — and the cycle collapses.
Corporate profits might surge initially. Companies that adopt AI early will slash costs and dominate competitors. But as more firms follow suit and fewer humans earn wages, the customer base erodes. You can’t sell products to people who have no income. It’s a paradox that Henry Ford understood a century ago when he paid his assembly line workers enough to buy the cars they built. AI doesn’t buy cars.
Gawdat sees this as a structural flaw, not a temporary dislocation. He argues that without massive intervention — wealth redistribution, new ownership models, or some form of universal stake in AI-generated productivity — the concentration of wealth will accelerate to levels that make today’s inequality look modest.
The numbers already point in that direction. The companies leading in AI development — Microsoft, Google, Meta, Amazon, Apple, Nvidia — have a combined market capitalization exceeding $15 trillion. Their workforces, while well-compensated, represent a tiny fraction of the global labor market. As these firms capture more economic value through AI, the gap between capital owners and everyone else widens.
This is where Gawdat’s warnings intersect with a growing chorus of concern from economists, technologists, and even some business leaders. OpenAI CEO Sam Altman has discussed the need for something like “universal basic compute” — giving every person a share of AI processing power as a form of economic participation. Elon Musk has called AI “the most disruptive force in history.” Dario Amodei, CEO of Anthropic, has written extensively about both the promise and the existential risk.
But rhetoric hasn’t translated to policy. Not yet.
The U.S. Congress has held hearings. The European Union passed the AI Act, focused primarily on safety and transparency. China is investing heavily in AI while tightening state control over its deployment. None of these efforts address the macroeconomic disruption Gawdat describes. They’re regulating the technology. He’s talking about the economy that technology is about to reshape.
There’s a temporal mismatch that makes the problem worse. Democratic governments operate on two- to six-year election cycles. Corporate boards think in quarters. AI capabilities are doubling in months. The entities best positioned to act are moving slowest, while the technology accelerates regardless of whether anyone is ready.
Gawdat has proposed that nations begin planning now for a transition period — a decade or so during which work becomes increasingly optional and new systems of value distribution take hold. He acknowledges this sounds utopian. But he argues the alternative — letting market forces sort it out — would produce dystopian levels of inequality, social unrest, and political instability.
The Risk of Getting It Wrong
What makes Gawdat’s perspective particularly compelling is his refusal to fit neatly into either the techno-optimist or doomer camp. He’s not arguing that AI should be stopped. He doesn’t think it can be. He’s arguing that the political and economic response to AI is dangerously inadequate.
He draws a distinction between the technology itself and the systems surrounding it. AI, in his view, could enable an era of unprecedented abundance — solving climate change, curing diseases, eliminating poverty. But only if the gains are distributed. If they’re captured by a small number of corporations and their shareholders, the result isn’t abundance. It’s feudalism with better algorithms.
There’s historical precedent for this kind of failure. The Industrial Revolution generated enormous wealth, but the first several decades were marked by brutal working conditions, child labor, and massive inequality. It took unions, regulation, and political upheaval to distribute the gains more broadly. That process took roughly a century. Gawdat doesn’t think we have that kind of time.
His warnings carry particular resonance in the current moment. Tech layoffs have accelerated even as AI investment soars. Companies are explicitly citing AI as a reason for reducing headcount. IBM, Google, Meta, and Amazon have all made significant cuts while simultaneously pouring billions into AI infrastructure. The pattern is clear: invest in machines, divest from people.
And the technology is still in its early stages. GPT-4, Claude, Gemini — these are impressive but primitive compared to what’s coming. Gawdat points to the trajectory: each generation of AI is dramatically more capable than the last, and the intervals between generations are shrinking. What happens when AI can do not just 30 percent of a knowledge worker’s job but 90 percent? Or 100 percent?
Some economists argue new jobs will emerge — roles we can’t yet imagine, just as the internet created entire industries that didn’t exist in 1990. Gawdat acknowledges the possibility but considers it insufficient. Even if new jobs appear, they won’t appear fast enough or in sufficient numbers to absorb the displaced workforce. And many of the new roles will themselves be quickly automated.
So where does that leave us?
In Gawdat’s view, at a crossroads. One path leads to a world where AI-generated wealth is broadly shared, where work becomes a choice rather than a necessity, and where human creativity and connection take precedence over economic productivity. The other path leads to mass unemployment, social fragmentation, and a concentration of power that would make the robber barons of the Gilded Age look like amateurs.
The choice, he insists, is political. Not technological.
That’s the part that should keep executives and policymakers up at night. The technology is coming regardless. The question is whether the institutions meant to govern society can adapt quickly enough to prevent a catastrophe that’s no longer hypothetical — it’s on a schedule.
Gawdat puts the timeline at roughly 2030 to 2035 for the most severe economic disruptions to begin manifesting. That’s not a generation away. It’s a business cycle away. And for an industry accustomed to thinking in quarters, that should feel uncomfortably close.
The Former Google Exec Who Thinks AI Will Collapse Capitalism Within a Decade first appeared on Web and IT News.