Predicting AI job exposure
Everyone is focused on AI job displacement metrics. Nobody is talking about why those predictions are fundamentally broken.
Why it matters
Benedict Evans argues that attempts to quantify AI job exposure are methodologically flawed because job transformation is non-linear, interdependent, and unmeasurable. This challenges the investor and founder playbook of using job displacement as a market-sizing proxy.
The key facts
9 to knowJobs won't simply disappear — they will change in unpredictable ways
Cascading second-order effects make isolated job exposure scoring impossible
Work itself is difficult to measure and compare across roles and industries
Current LLM progress benchmarks don't map cleanly to real-world job displacement
Author: Benedict Evans
Core claim: AI job exposure analysis is 'mostly impossible'
Three stated reasons for prediction failure: job transformation unknowability, systemic second-order effects, work measurement limitations
Published May 2026 — suggests post-hype maturation of AI labor discourse
No quantitative data provided — opinion-driven contrarian take
Go to the source
Benedict Evansben-evans.com
Publisher excerpt: Many people would like to analyse which jobs, companies and industries are most exposed to AI, and assign scores, build charts, and map that against the progress of LLMs. I think this is mostly impossible: you don’t know how the jobs will change, you don’t know what else will change around this,…
