For years, the standard economic response to automation anxiety followed a predictable script: technological disruption has always created more jobs than it destroyed, and artificial intelligence will be no different. That narrative is fracturing. A comprehensive multi-university study surveying 69 economists alongside 52 artificial intelligence researchers and 38 superforecasters—specialists trained in long-term prediction—has reached near-consensus on a more sobering conclusion: rapid AI advancement correlates directly with net job loss in the near term.
This represents a significant epistemic shift within the economics profession. The historical precedent of the Industrial Revolution and subsequent technological waves did eventually yield employment gains, but the lag between displacement and retraining was measured in decades, not quarters. Current AI deployment cycles operate at a fundamentally different velocity. Large language models and generative systems are already demonstrating measurable productivity gains across white-collar work, from legal research to software engineering to financial analysis. Unlike previous automation waves that primarily affected manufacturing and routine labor, this wave targets knowledge work—traditionally the most resilient sector against displacement. The speed of capability improvement, combined with capital's incentive to substitute labor immediately upon availability, creates structural challenges that historical comparisons may not adequately capture.
What makes this consensus notable is its breadth across disciplinary boundaries. Economists bring empirical rigor but have historically underestimated technological disruption timelines. AI experts understand deployment realities and capability trajectories more granularly than macro observers. Superforecasters, selected for predictive accuracy across diverse domains, represent a meta-level reality check on expert overconfidence. Their alignment suggests this isn't ideological positioning but rather convergence on observable trajectory data.
The acknowledgment carries important caveats. Job loss and employment destruction are not identical; labor market transitions, retraining programs, and policy interventions remain levers for mitigation. The study essentially confirms that without deliberate policy response, AI-driven productivity gains will compress labor demand faster than historical precedent would suggest. This creates space for serious discussions around education infrastructure, social safety nets, and potential structural reforms like income support mechanisms—conversations that seemed premature when the dominant narrative held that concerns were overblown. Whether policymakers can implement solutions at the speed and scale AI advancement requires remains the critical open question.