A new economic model reveals that AI automation creates a market failure where rational firms collectively harm themselves by over-automating. The research demonstrates that when AI displaces workers faster than economic absorption, it erodes consumer demand that businesses depend on, creating a destructive feedback loop.

In a competitive task-based framework, demand externalities trap companies in an automation arms race that exceeds socially optimal levels. The study finds that traditional policy responses - including wage adjustments, universal basic income, upskilling programs, and even worker equity participation - cannot correct this market failure. More competition and advanced AI actually worsen the problem.

Only a Pigouvian automation tax can address the root cause by internalizing the negative externality firms impose on each other. The findings suggest policymakers must look beyond traditional approaches like capital income taxes or Coasian bargaining. The research indicates that addressing AI's labor market impact requires tackling the competitive dynamics that drive excessive automation, not just managing its consequences.