The AI Labour Market: A Prediction Market Exists for This
The question "will AI replace X% of current jobs by [date]?" is genuinely uncertain enough to be a meaningful prediction market. Economists disagree significantly: McKinsey Global Institute estimates 15% of current work is automatable in the next decade; others suggest 30-40%; AI researchers like Dario Amodei speak of AI potentially performing the work of "almost all cognitive workers" within a few years. That range of expert opinion is exactly what prediction markets are designed to price.
How to Think About This Prediction Market
- →The "task" vs "job" distinction matters: AI automates specific tasks within jobs, not usually entire jobs. Prediction markets need to be carefully scoped.
- →Sector-specific timelines: coding, customer service, and data analysis face much shorter disruption timelines than plumbing or nursing
- →Labour market adaptation: past automation (ATMs didn't eliminate bank tellers) shows the economy creates new roles even as tasks are automated
- →Regulatory friction: employment law, liability frameworks, and union agreements significantly extend the timeline between technical capability and economic deployment
- →Human preference: some tasks will remain human-performed even after automation is possible, because customers or regulations prefer human interaction
"The real prediction market question isn't "will AI replace jobs?" — it's "in which sectors, on what timeline, with what regulatory response?" Those are tractable questions."
Boromarket's AI labour market prediction series includes sector-specific markets (legal, finance, creative, technical) with different resolution dates and conditions.