No forecasts. Only what has already happened — every line from a public, checkable, still-updating source.
Last updated: 2026-10-24
Read carefully: these show pressure, not collapse. No evidence yet supports economy-wide AI job loss. Reading them as doom repeats the error we criticise in others.
Stanford’s August 2026 revision reports a 19% employment shortfall for workers aged 22–25 in highly exposed occupations, relative to a counterfactual tracking less-exposed peers, as of June 2026. The comparable July 2025 vintage was 15%. Adjustment appears mainly through reduced hiring. These are descriptive patterns, not causal estimates of AI’s effects.
Share of occupations where AI covers a quarter of tasks or more
Under review
Under review
The previously displayed 49% occupational task-coverage figure could not be verified in the cited report text. Its value and trend are withheld pending verification. Claude usage coverage is not an automation or unemployment rate.
Share of AI use that automates rather than augments
Under review
Under review
The previously displayed automation (45%) and augmentation (52%) figures could not be verified in the cited report text. Values and trends are withheld pending verification. Usage patterns alone do not establish job displacement.
The IMF estimates about 60% of jobs in advanced economies are exposed to AI, against 40% in emerging markets and 26% in low-income countries — higher in rich countries because cognitive roles are denser there. The IMF further estimates roughly half of exposed jobs in advanced economies face adverse effects: lower demand, lower wages, less hiring. This is a structural estimate rather than a live indicator, so it does not move month to month.
This line is here because it damages our own case, and leaving it out would make us dishonest. The US ran a $1.8 trillion deficit in FY2025, with net interest payments passing $1 trillion for the first time. So at the same moment AI employment pressure is rising, the fiscal room to respond to it is shrinking, not growing. Any UBI proposal has to be designed against this fact rather than around it.
China's population stood at 1,404.89 million at the end of 2025, falling for several consecutive years. The implication for UBI runs both ways: a smaller denominator lowers the total cost of paying everyone, but the working-age population is shrinking alongside it, so the taxpayers carrying that cost are fewer too. Reading this as simply good or bad news for UBI is lazy.
Every reading carries an "as of" date, which is when a human opened the source and checked it — not an automated scrape timestamp. The repository includes a script (npm run update:pressure) that diffs the source pages and lists which indicators need re-checking. We do not auto-extract figures: regexing a percentage out of a web page and publishing it as fact is the most common way sites like this go wrong.
Why fiscal constraints are included
Income-security options must be considered alongside employment changes and fiscal constraints. Indicators are not selected to favor or oppose UBI. Evidence limiting policy options belongs here too. Readings under review are withheld and should not be used to infer a trend.