The 2028 Global Intelligence Crisis: What If AI Works Too Well?

In 1938, Orson Welles broadcast a radio adaptation of The War of the Worlds so realistic that thousands of listeners believed Martians were actually invading Earth. It didn’t need to be true: it just had to sound credible enough to trigger panic.

Almost ninety years later, a document published on Substack pulled off something remarkably similar on Wall Street. Citrini Research published The 2028 Global Intelligence Crisis, a piece of financial science fiction that triggered 4% to 6% drops in shares of Uber, Visa, Mastercard, and DoorDash in a single trading session. It wasn’t a negative earnings report or a Fed decision. It was a fictional narrative so well-crafted that markets took it seriously.

Its premise is unsettling: what happens when AI works too well?

What the Citrini report argues

The document presents itself as a fictional macro memo written from June 2028. In this hypothetical scenario, the S&P 500 has fallen 38% from its highs, U.S. unemployment has surged to 10.2%, and the economy is trapped in a deflationary spiral.

The central thesis isn’t that AI fails - it’s the opposite: intelligence, historically the most valuable and scarce resource, becomes a cheap commodity. When that happens, the value of skilled human labor collapses.

The report describes what it calls the “human intelligence displacement spiral”: companies replace white-collar workers with AI agents, laid-off workers spend less, consumption falls, companies automate further to compensate, and the cycle accelerates.

Ghost GDP: the economy grows, but nobody feels it

One of the report’s most powerful concepts is “Ghost GDP.” The economy produces more thanks to AI, productivity soars, nominal GDP grows, but that wealth doesn’t circulate through the real economy. It concentrates among compute owners while real wages stagnate or decline.

In Citrini’s words: “a single GPU cluster in North Dakota generates the output previously attributed to 10,000 white-collar workers in midtown Manhattan.” The problem isn’t production - it’s that machines don’t buy homes, don’t pay mortgages, and don’t spend on consumer goods.

In their scenario, labor’s share of GDP falls to 46%, and the $13 trillion mortgage market comes under stress because prime borrowers, skilled professionals, no longer have the income that justified their loans.

Three sectors in the line of fire

The report identifies three particularly vulnerable industries:

1. Software as a Service (SaaS)

With tools like Claude Code or Codex, a competent developer can replicate the core functionality of a mid-market SaaS product in weeks. Citrini imagines CIOs starting to ask: “what if we just built this ourselves?” The result: renewals with 30% discounts, collapse of the SaaS long tail (Zapier, Monday.com, Asana), and an all-out price war between incumbents and agile startups.

There’s also a devastating reflexive effect: when ServiceNow’s clients cut 15% of their workforce, they cancel 15% of their licenses. The same automation improving their clients’ margins destroys their own revenue base.

2. Intermediation platforms

Uber, DoorDash, and similar models depend on human friction: the lack of time to compare prices, loyalty born of convenience, inertia. When autonomous AI agents compare options in real time and always choose the cheapest, loyalty to any single app evaporates. Customer lifetime value, the metric on which the entire subscription economy was built, plummets.

3. Payment networks

Visa, Mastercard, and American Express charge fees on every transaction. If AI agents prioritize faster and cheaper systems (stablecoins, direct transfers, new payment rails), fee revenue for these networks could shrink dramatically.

The critics: not so fast

The report hasn’t gone unanswered. Citadel Securities, Ken Griffin’s firm, published a sharp counter-analysis dismantling several pillars of the thesis.

The data tells a different story

Citadel points out that demand for software engineers is actually rising, not falling, up 11% year over year in early 2026, according to Indeed data. Daily generative AI adoption at work remains “unexpectedly stable,” with no evidence of imminent displacement. And new business formation in the U.S. is expanding, not contracting.

The infinite recursion fallacy

Citrini’s core error, according to Citadel, is conflating recursive technology with recursive economic adoption. The fact that AI can write code to improve itself doesn’t mean its integration into the economy compounds infinitely and instantaneously.

Technology diffusion has historically followed an S-curve: slow early adoption, acceleration as costs drop, and a plateau as saturation sets in. Moreover, there are physical constraints Citrini ignores: energy and compute capacity. If automation expanded at the pace Citrini fears, compute demand would push its marginal cost above the cost of human labor for many tasks, creating a natural brake.

Noah Smith: a scary bedtime story

Economist Noah Smith was even more direct, calling the report a “scary bedtime story”, well-written but lacking solid empirical foundations. His argument: the narrative is captivating precisely because it appeals to real fears, but it ignores the adjustment mechanisms that market economies have historically demonstrated.

What this means for businesses

Beyond the bull-vs-bear debate, the Citrini report has put questions on the table that every company should be asking:

Capability diversification

If your business model depends on human intelligence being scarce and expensive, you need a plan B. AI isn’t going away, and its cost will keep dropping.

Reskilling as investment, not expense

Companies that invest in reskilling their teams, not just to use AI, but to work with AI, will have a competitive edge over those that simply cut headcount.

AI as ally, not replacement

The most likely scenario isn’t Citrini’s apocalypse or unlimited bull optimism. It’s a middle ground where AI amplifies human productivity, but requires new models for distributing the generated wealth.

Building for volatility

Regardless of who’s right, the uncertainty is real. Companies that build resilience (revenue diversification, operational flexibility, a culture of adaptation) will be better positioned for any scenario.

Conclusion

The Citrini report isn’t a prediction. It’s a thought experiment. But its power lies in articulating a legitimate fear: that humanity’s greatest technological achievement could, paradoxically, weaken the economic structures on which our society is built.

The answer isn’t to slow down AI. It’s to prepare for a world where intelligence is abundant and cheap, and to redesign economic systems so that the benefits of that abundance reach everyone, not just the owners of compute.

Is your company ready for that scenario? At My Tech Plan, we help organizations integrate AI strategically, without losing sight of the human factor.