Artificial intelligence is being promoted as the technology that will “change everything.” Yet while a handful of firms are profiting enormously, a different question deserves attention:
Is AI accelerating the economy — or merely masking its slowdown?
Across headlines, AI is credited with transforming medicine, finance, logistics, commerce, and productivity. Yet among many working professionals, there is a sense that real wages have struggled to keep pace with rising living costs. At the same time, the loudest optimism often comes from sectors most financially invested in the AI narrative.
This raises an uncomfortable question: Has AI become a true engine of prosperity — or a financial life-support system?
The Mirage of Growth
Recent data suggests a substantial portion of U.S. GDP growth may be driven not by rising productivity, but by AI-related infrastructure spending — especially data centers.
A study from S&P Global found that in Q2 of 2025, data center construction alone added 0.5% to U.S. GDP. That is a historic figure. But what happens if this spending slows?
Are we seeing real economic expansion — or a short-term stimulus disguised as innovation?
This pattern has precedents. Before the 2008 housing collapse in Ireland — and in the United States the same year — construction boomed, GDP rose, and optimism became mandatory. Economies looked healthy on paper. Fragility was already taking root.
Today, the boom is not in bricks and concrete — but in silicon, servers, and expectation.
The Productivity Paradox
AI has been advertised as a labor-saving miracle. Yet many businesses report a different reality: “work slop” — material produced by AI that looks polished but must be repaired, rewritten, or verified by humans. Time is not saved — it is quietly relocated. Studies suggest the same paradox:
According to media coverage, MIT found that 95% of corporate AI pilot programs showed no measurable ROI.
MIT Sloan research indicated that AI adoption often lowers productivity initially, and improvements (it is claimed) only occur after major organizational restructuring.
Even McKinsey— one of AI’s strongest advocates — warns that “piloting AI is easy, but creating value is hard.”
If AI has not reduced human labor — has it merely concealed it?
If AI is not transforming work — is it simply rebranding it?
In addition, AI may appear efficient, but it operates strictly within the limits of its training data: it can replicate mistakes, miss what humans would notice, and reinforce boundaries that present an “approved” or “consensus version” of reality rather than reality itself.
From Innovation to Infrastructure
AI is now being treated less like a tool — and more like a necessity. Once a technology becomes classified as essential infrastructure, funding shifts from private risk to public obligation.
McKinsey estimates that over $6.7 trillion may be spent on AI and computing infrastructure by 2030 — an investment scale usually seen in wartime. Yet what is being built, and will it generate tangible returns for ordinary citizens?
Some analysts warn that sectors of the AI economy resemble a circular investment loop: cloud and chip firms invest in AI startups that then purchase computing services from the very firms that funded them. Demand becomes proof of viability — even if value remains unproven.
As explored in the books The Debt Machine and Demonic Economics, these cycles often create an illusion of growth — until the public is left to carry the cost. At that point, AI stops behaving like innovation — and starts behaving like debt.
The Genesis Mission — A Turning Point
In November 2025, the U.S. government issued an Executive Order launching the “Genesis Mission,” a national effort to build an integrated AI platform — combining DOE national-lab supercomputers, federal supercomputers, decades of federally funded scientific datasets, and cooperation with private-sector AI and academic partners — to accelerate scientific research and innovation.
This does not guarantee bailouts — but it creates the conditions under which AI firms may become “too strategic to fail.” Once a technology becomes part of national policy, failure shifts from financial risk to political risk.
We may be witnessing the transformation of AI from speculative investment into publicly underwritten infrastructure.
Who Carries the Risk?
The deeper question is not merely what AI can or can’t do — but who must carry the cost if it fails.
Large U.S. retirement funds and passive index portfolios are now heavily exposed to AI-dependent giants such as Nvidia, Amazon, Microsoft, Google, and Tesla. On the debt side, private credit and data-center financing tied to AI infrastructure are steadily entering bond portfolios.
This means the AI boom is not just an investment trend.
It may already be embedded in retirement accounts — without citizens’ knowledge.
Are we potentially witnessing a new socialization of private risk and debt — similar to the aftermath of the 2008 housing collapse — with the burden once again shifting to the public?
Questions That Demand Scrutiny
Is AI increasing productivity — or disguising its decline, and creating new layers of hidden labor?
Are data centers driving prosperity — or propping up GDP on paper?
Are citizens knowingly investing in AI — or are their pensions investing for them?
Is AI creating value — or absorbing subsidies and debt?
When a technology is credited with growth, shielded from failure, and treated as essential before its value is proven — it may no longer be innovation. It may be obligation.
Conclusion
As I wrote in the book Staying Human in the Age of AI we should not allow AI to overshadow human thought. There is still time for AI to deliver genuine breakthroughs. But at this moment, belief is moving faster than evidence — and optimism is cheapest when it costs the public nothing.
The cost of failure will not fall on Silicon Valley.
It may fall on pensioners, savers, and future generations.
Once a technology becomes too strategic to fail — it may no longer operate as innovation — but as a public obligation that few ever asked for.
Mark Keenan is a former United Nations technical expert who writes on the intersection of science, finance, and public policy. He is the author of The Debt Machine, Demonic Economics, Climate CO₂ Hoax, and Staying Human in the Age of AI. He publishes at markgerardkeenan.substack.com
