Could AI Crash the Economy in 2 Years? One Research Firm Says Yes.

• Citrini's 'doom loop' scenario describes companies increasingly investing in AI to cut costs, leading to white-collar layoffs that reduce consumer spending and create 'Ghost GDP' - economic output that never circulates through the real economy
• The report caused immediate market reactions, with IBM stock falling after AI startup Anthropic announced a COBOL modernization tool that threatens one of IBM's legacy assets
• Economists have challenged the scenario citing Say's Law (supply creates its own demand) and arguing AI could increase overall employment by enhancing worker productivity rather than replacing workers entirely

Summary: AI Economic Crash Warning by Citrini Research

Key Points

Investment research firm Citrini Research released a report titled "The 2028 Global Intelligence Crisis," outlining a scenario where AI-driven job displacement could push U.S. unemployment above 10% and trigger an economic crisis within two years (by 2028).

Core Thesis

The report describes a "doom loop" where:

• Improving, cheaper AI enables mass white-collar layoffs
• Companies reinvest savings into more AI capabilities
• Displaced workers reduce spending, decreasing aggregate demand
• Margin pressures force companies to invest further in AI
• This creates a "negative feedback loop with no natural brake"

The report introduces the concept of "Ghost GDP"—economic output visible in GDP and corporate profits but not circulating through the real economy.

Market Impact

The S&P 500 fell 1% on Monday, February 23, 2026, following the report's release. Major tech stocks showed mixed performance, with NVDA down 4.4% and MSFT down 2.2%.

IBM stock declined 0.74% to $240.21 after AI startup Anthropic announced its Claude Code tool could modernize COBOL coding language, threatening IBM's legacy assets.

Expert Response

Economists have criticized the report's assumptions as "wildly speculative." Counter-arguments include Say's Law (new supply creates its own demand) and the possibility AI could increase overall employment through productivity tools.

Important Note: Citrini's authors emphasized this is a scenario, not a prediction.