The Risks of AI Doom Loops, Ghost GDP, and Automation in the Workplace

A “thought experiment turned viral essay” has posed a concerning question: what happens when AI significantly boosts profits but reshapes the foundations of demand? “The 2028 Global Intelligence Crisis,” co-authored by AI entrepreneurs James Van Geelen, founder of Citrini Research, and Alap Shah, imagines a near-future where real wage growth is negatively impacted by AI, while the economic gains for those who own computing infrastructure increases dramatically.

Since its publication in February 2026, the essay has prompted a wide range of conversations about the future of AI's economic impact—with business leaders, economists, and technologists offering starkly different views on what automation and a potential AI doom loop could mean for workers and income growth.

What Is Ghost GDP?

Central to the essay is a concept the authors call “Ghost GDP,” where financial gains appear in national accounts but never actually circulate through the real economy. In this scenario, AI agents replace software engineers, analysts, middle managers, and other white-collar workers. Corporate profits surge. But, unlike people, the AI machines that now generate the work output spend no money in this economy.

As jobs become scarce, displaced workers cut back more and more on spending, and consumer demand shrinks, which pushes companies to automate further and lay off more workers. Van Geelen and Shah call this self-reinforcing cycle the "human intelligence displacement spiral,” or an AI doom loop. It is a negative feedback loop with no natural brakes. The authors suggest that it could produce an S&P 500 stock market index plunge of 38%, unemployment above 10%, and a chaotic mortgage market in our near future.

Economists Push Back

Not everyone believes the AI doom loop theory holds up under scrutiny. Economists have called the essay "allegorical," arguing it ignores a basic principle—that production typically generates income that goes to someone who will spend or invest it. Citadel Securities published a point-by-point rebuttal, contending that energy costs and computing capacity will naturally halt any runaway AI replacement spiral before it reaches catastrophic levels.

Current labor data supports these skeptics, citing that white-collar job postings are stable. While Deutsche Bank projects 92 million roles will be eliminated by 2030, they also suggest that 170 million new jobs will be created. The essay's authors themselves framed their work as a thought exercise, not a play-by-play prediction per se, although their perspective warrants consideration.

Where Workplace Automation Makes Sense

Organizations that over-automate face risks at every level. Many CEOs are not aiming to use AI as a blunt method to reduce headcount. The more deliberate leaders are instead using AI to redesign work by identifying which tasks are genuinely automatable and which still require or benefit the most from human judgment.

Automation delivers real value in repetitive, rules-based work, such as claims processing, compliance checks, fraud detection, and customer routing. For example, the IRS now uses AI to flag audit risks and triage taxpayer inquiries. Companies like IBM and Cognizant are simultaneously expanding junior hiring by 20% or more, using AI to enhance early-career workers rather than replace them entirely. Tanmai Gopal of PromptQL estimates that 70% of tasks cannot be fully automated because they depend on a level of human management that current AI models cannot capture reliably.

And when institutional knowledge is stripped out in the name of efficiency, companies lose the adaptability they need when conditions change. Experts point to Australia's Robodebt scheme, where automated debt recovery wrongly targeted hundreds of thousands of welfare recipients, as a cautionary tale of what happens when AI replaces human judgment without meaningful accountability.

Job Security for Those Who Manage AI

Whether or not the Ghost GDP scenario materializes, one outcome is already clear: the gap between workers who can direct AI systems and those who cannot is widening fast. When Jack Dorsey, co-founder of Twitter (now X) and Square (now Block), announced that Block was cutting 40% of its workforce because "intelligence tools have changed what it means to build and run a company," it signaled a shift already underway. Workers who understand how to deploy, oversee, and critically evaluate AI systems will not just survive this transition, they will lead it.

The stakes raised by the Ghost GDP debate underscore how much of the world depends on AI professionals who understand not just how these systems work, but how they reshape organizations and society. Capitol Technology University's Master of Research in Artificial Intelligence prepares graduates to work at the frontlines of this evolving technology. Through rigorous research, applied AI methods, and a curriculum grounded in both technical depth and ethical responsibility, the program develops specialists who can build AI systems responsibly, assess their societal impact, and help organizations navigate ethical transitions in the workplace.

Explore what a degree from Capitol Tech can do for you! To learn more, contact our Admissions team or request more information. 

Written by Jordan Ford 
Edited by Erica Decker