How Could an AI Boom Trigger a Wider Economic Crash?

This complexity raises the danger of rapid contagion within the sector and across the broader economy, mirroring the interconnected vulnerabilities that triggered the 2007-2008 financial crisis. Infrastructure spending on AI has already strained supply chains across a range of goods, contributing to broader economic pressures, such as labor shortages and capital allocation challenges in non-AI industries.

The paper draws lessons from past crises, noting that the 2008 government's bailout approaches were costly and not without shortcomings, and argues that Congress should not bail out AI firms, related technology companies, or affected financial firms, although any financial relief should come with public-benefit commitments, such as public equity stakes.

Assessing Systemic AI Risks

The paper draws on economic history, financial regulation, AI technological development, and labor market studies. The author surveys financial filings, market analyses, and infrastructure investment data to diagnose the systemic risks posed by AI-related financial arrangements. The paper reviews policy responses to previous crashes, including legislation like the Dodd-Frank Act, to inform recommendations tuned to the distinctive characteristics of AI markets. Additionally, the paper draws on contemporary commentary and expert assessments on AI's impact on employment and the social and political economy implications of AI infrastructure deployment to propose policy interventions for sustainable AI innovation, worker protection, and regulatory oversight. Rather than reporting a formal empirical methods section, the paper takes a forward-looking policy approach aimed at formulating options before a crisis unfolds.

Financial Fragility and Policy Gaps

The paper identifies "extreme financialization" in the AI sector, marked by circular equity financing among top companies such as Nvidia, Microsoft, and OpenAI, which creates fragile interdependencies that magnify the risk that problems in one entity cascade to others. The extensive, sometimes hidden, use of debt instruments and varying government subsidies further complicates transparency and risk evaluation. This financial complexity risks precipitating an economy-wide crash beyond a mere "bubble burst," with cascading effects causing broad business failures, sharp stock market declines, reduced lending, and labor market dislocations. The article argues that following such a crash, policymakers must resist repeating past mistakes such as hasty, narrow legislation and ineffective bailouts. Instead, it recommends structural reforms, including the curtailment of risky financial engineering practices, prosecution of financial fraud linked to AI investments, and repurposing stranded physical infrastructure (like data centers) to serve public interests through a publicly accessible cloud platform.

Worker protection emerges as a key focus area, given the uncertain impact of AI on employment. The article notes contrasting CEO views on AI’s employment effects and empirical studies showing mixed evidence on job displacement. Anticipating increased unemployment from both an AI-driven productivity shock and the aftermath of a crash, policies such as expanded and more accessible unemployment insurance, creation of a digital works progress administration, and strict limits on workplace surveillance are recommended to mitigate employment harm. Moreover, the paper stresses the necessity of maintaining robust publicly funded AI research and development for socially beneficial use cases that private firms might abandon post-crash.

Recognizing the unique challenges of AI as a digital utility, the report advocates separating algorithmic control from data center ownership through a "Glass-Steagall"- style regulatory separation. It also proposes a new federal regulatory agency equipped to comprehensively oversee AI and digital markets, encompassing privacy, competition, and transparency concerns, and to be accompanied by sector-specific regulations akin to those for traditional utilities. The paper also explicitly calls for banning surveillance-based business models, including surveillance advertising and surveillance prices and wages, as part of its broader reform agenda.

Toward a More Stable AI Future

The paper concludes by emphasizing the importance of proactive policymaking to confront the vulnerabilities and inequalities emerging from AI-related financial and economic dynamics. It warns against delaying reforms until after a crisis, when political will concentrates, but options may be limited or inadequately formed. Implementing the proposed bold reforms now may not only reduce the likelihood or severity of a crash but also guide the AI industry toward innovations aligned with public good, worker protection, and economic stability. In conclusion, the paper presents a comprehensive framework for legislative and regulatory intervention aimed at fostering a resilient, transparent, and socially responsible AI ecosystem before the next major disruption occurs.

*Important notice: SSRN publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as definitive, used to guide development decisions, or treated as established information in the field of artificial intelligence research.