Raghuram Rajan Proposes AI Token Tax to Prevent 'Jobocalypse'

Economist and former Reserve Bank of India governor Raghuram Rajan has proposed a targeted tax on AI tokens used by corporations to mitigate the risk of a widespread "jobocalypse." The proposal seeks to neutralize the financial incentives that currently drive firms to replace human employees with automated systems.

In a note titled "How Corporations Can Mitigate an AI Jobocalypse," Rajan argues that current fiscal structures create an unfair advantage for automation. He points to a specific tax imbalance in the United States, where corporations are required to make social security payments for human workers but face no such obligations for the AI systems that replace them. To level the playing field, Rajan suggests a carefully calibrated levy on the AI tokens—the basic units of text or data processed by large language models—that firms use to power their operations. This mechanism would simultaneously discourage purely cost-driven displacement and bolster government revenue.

The Economic Context

The proposal arrives as generative AI continues to integrate into corporate workflows, sparking fears of systemic unemployment. Rajan acknowledges that while AI-related job displacement is inevitable, the speed and scale of the transition remain uncertain. He notes that no one yet knows how fast the process will proceed, how far it will go, or which specific sectors will be hit hardest.

However, Rajan also references the "Jevons effect," an economic theory suggesting that as technology increases productivity and lowers the cost of a service, the overall demand for that service may actually increase. In this scenario, AI could potentially boost total employment by expanding the market for goods and services, provided the transition is managed correctly.

Why It Matters

This shift in strategy moves the conversation from passive prediction of job losses to the creation of active fiscal guardrails. By taxing the "inputs" of AI and subsidizing the "inputs" of human capital, the goal is to ensure that technological progress results in high-quality employment rather than mass displacement.

To support this transition, Rajan proposes pairing the token tax with tax credits for corporations that invest in their workforce. These credits would be awarded to firms that provide additional training to employees, with the financial benefits explicitly linked to retention. Specifically, the proposal suggests that one-third of the credit's value be usable for each year a worker remains employed at the company.

What's Next

While the proposal provides a theoretical framework for government intervention, its implementation would require significant international coordination to prevent firms from shifting AI workloads to low-tax jurisdictions.

There is already some evidence of a corporate shift toward valuing human capital; Rajan notes that the share of US Fortune 150 CEOs mentioning employee development in shareholder letters rose from approximately 20% in 2008 to 44% in 2023. Whether this trend is sufficient to offset the efficiency of AI, or if a formal token tax becomes necessary to protect the labor market, remains the central question for policymakers.