Is AI too successful, causing an economic crisis? Institutions simulate 2028: Unemployment rate exceeds 10%, S&P plunges 38%

While the market remains immersed in the productivity revolution brought by AI and corporate stock prices continue to hit new highs, research firm Citrini Research presents a starkly different scenario: if artificial intelligence develops smoothly with continuous exponential improvements, the current economy centered on white-collar income and consumption could face systemic upheaval. The report hypothesizes that by 2028, the U.S. unemployment rate could rise above 10%, the S&P 500 could decline by 38% from its 2026 peak, and a chain reaction called the “Global Intelligence Crisis” is unfolding.

Is AI an opportunity or a threat? Citrini predicts high unemployment and a stock market crash in 2028

Citrini Research’s report is written in a future retrospective style, assuming it is mid-2028 and looking back at this AI-driven economic shock:

This morning’s unemployment rate was announced at 10.2%, 0.3% higher than expected. The market reacted with a 2% decline, bringing the S&P 500’s cumulative drop since its October 2026 high to 38%.

In this hypothetical scenario, 2026 was a feverish market period. The S&P 500 approached 8,000 points, and the Nasdaq broke through 30,000. Companies drastically cut white-collar jobs, replacing them with AI, which boosted profit margins and beat earnings expectations, driving stock prices higher. On the surface, nominal GDP maintained mid-to-high single-digit growth, and productivity data hit decades-long highs.

However, the report points out that it is precisely in this “AI revolution exceeding expectations” scenario that potential risks begin to accumulate.

Ghost GDP? Disconnection between labor and consumption markets

Citrini introduces the concept of “Ghost GDP,” referring to a situation where output and productivity continue to grow in statistical data, but the benefits are not translated into household disposable income and consumption capacity.

In this hypothetical scenario, AI agents can autonomously handle R&D, programming, and decision-making tasks for extended periods, leading to a sustained decline in demand for white-collar labor. White-collar workers’ incomes are squeezed; even if some shift to service or gig economy jobs, their wages are significantly lower than their previous positions.

As white-collar workers are forced into service and gig markets, the surge in supply will exert broader wage pressures on the labor market.

The problem is that the U.S. consumption structure heavily relies on high-income groups. The report notes that the top 10% contribute over 50% of consumer spending. If this group’s income is cut, even a relatively limited unemployment rate could have multiple times the impact on overall consumption.

Against this backdrop, economic growth data still shows positive figures, but real consumption momentum begins to weaken, leading to a decoupling of output and demand.

Non-cyclical recession: AI cycles without natural brakes

The report emphasizes that this hypothetical crisis differs from traditional economic cycles. Typical recessions often have self-correcting mechanisms, such as destocking or renewed investment and consumption following interest rate cuts. However, AI-driven adjustments resemble large-scale “industry structural replacements.”

The cycle operates as follows:

AI capability improvements → companies cut white-collar jobs → displaced workers reduce consumption → companies face demand pressure → to maintain profits, further invest in AI → AI again boosts efficiency and reduces human labor.

Citrini states, “This feedback loop can be described as a ‘no natural brake’ cycle.”

Meanwhile, AI agents are rewriting payment and intermediary models. Stablecoins and on-chain settlements reduce transaction costs; traditional credit card and payment network fees are compressed; industries relying on information friction and brand loyalty, such as travel, insurance, real estate, and consulting, face margin pressures due to AI automatic price comparison and decision-making. These changes will further weaken profit sources in some financial and service sectors.

(AI Agent era? Underlying protocols and payment systems for autonomous machine trading)

Private credit and life insurance chains: profit flywheel becomes a pressure source

On the financial front, the report focuses on risks in private credit and life insurance asset allocations. Over the past decade, the size of the U.S. private credit market has grown rapidly, with a significant portion invested in tech and SaaS companies, betting on their stable subscription revenue (ARR).

Citrini hypothesizes that as AI replaces some software and service functions, related companies’ revenues will be revised downward, default rates will rise, and credit ratings will be downgraded. Life insurers holding related assets may face increased capital requirements, potentially forcing them to sell less liquid assets, exacerbating market volatility.

In this scenario, the profit flywheel is not entirely stable, especially as the linked policy and annuity liabilities, under regulatory adjustments to risk weights, could quickly impose capital pressures.

(Anthropic launches AI safety tool Claude Code Security, causing several cybersecurity stocks to plummet)

2028 Mortgage Crisis: Changing income expectations as the main cause

Another key assumption in the report is potential stress in the mortgage market. Unlike the subprime crisis of 2008, this issue does not stem from low-credit borrowers or poor loan quality but from changes in borrowers’ future income expectations.

Citrini assumes that high-income white-collar workers in tech and finance hubs will face unemployment or pay cuts. Initially, they might rely on savings, home equity lines of credit (HELOC), or retirement accounts to meet mortgage payments, but over time, delinquency rates will rise. If housing prices fall by 8% to 11% annually, concentrated in tech hubs like San Francisco, Seattle, New York, and Austin, it could further impact the financial system.

The report indicates that in an extreme scenario, if cracks appear in the mortgage market, the stock market could ultimately fall by nearly 57%, similar to the 2008 global financial crisis, with the S&P 500 dropping to around 3,500 points.

Race against time: can policy responses come in time?

The challenge for governments is that tax revenue is highly dependent on labor income. As AI boosts productivity, the share of labor income in GDP will decline: “U.S. federal revenue could be more than 10% below expectations, while expenditure needs increase due to unemployment and transfer payments.”

The report suggests initial policy ideas, including taxing AI inference compute power or establishing a “shared AI dividend mechanism” similar to sovereign wealth funds, but political disagreements and policy lagging behind reality remain risks.

Repricing human intelligence: how should investors respond?

Most importantly, “human intelligence,” compared to capital and natural resources, is arguably the key scarce factor driving economic growth, and its value is facing revaluation: “This revaluation process will be painful and chaotic, but it does not mean collapse; economic structures and financial systems will find new equilibrium points on their own.”

Citrini emphasizes that these predictions are not certain to happen, but many aspects will ultimately face AI’s stress testing. The question is whether current economic and financial systems can adapt in time:

We are still in 2026, near all-time market highs, and the non-cyclical recession has not fully unfolded. Mechanical intelligence will continue to accelerate, and the premium on human intelligence is shrinking. As investors, we still have time to assess how much of our portfolios cannot withstand the next decade.

Did AI’s success trigger an economic crisis? The institution’s projection of 2028: unemployment over 10%, S&P down 38% first appeared on Chain News ABMedia.