Minerval
← claim pagemap viewclick a claim to focus on it · hover to preview · ⌫ back
trailAI is more likely to complement most jobs than to substitute for them entirely
Partial task exposure· for
Roughly two-thirds of U.S. occupations are exposed to some degree of automation by AI
Task-based analysis of occupational databases can meaningfully estimate occupations' exposure to AI automation
About 80 percent of US workers could have at least 10 percent of their tasks affected by LLMs.
About 60% of jobs in advanced economies are exposed to AI
For AI-exposed occupations, roughly a quarter to half of their workload could be automated
atomic
Historical precedent· for
Historically, automation technologies have created more jobs than they have eliminated
Aggregate employment in industrialized economies has grown over the long run despite widespread automation
Most U.S. workers today are employed in occupations that did not exist in 1940
Automation-driven productivity gains raise demand for labor in new and existing tasks
Since the 1980s, automation in the United States has displaced labor faster than it has created new tasks for workers
+1
Early adoption evidence· for
Generative AI adoption has so far had minimal effects on aggregate employment and wages
Capability trajectory and early displacement· against
AI capabilities will advance to the point of fully automating most occupations
US early-career workers aged 22-25 in the most AI-exposed occupations saw a 16% relative employment decline since generative AI's widespread adoption
The observed AI-related employment decline is robust to excluding technology firms and remote-work-amenable occupations
Employment for less AI-exposed workers and more experienced workers in AI-exposed occupations has stayed stable or grown
ADP payroll data are broadly representative of overall US labor-market trends
Hiring for AI-exposed occupations began declining before ChatGPT's release in November 2022
A claim that one thing brings about another, not merely that the two go together.constitutionclaim page ↗︎
AI is more likely to complement most jobs than to substitute for them entirely
Evidence favors the claim, but the chain is incomplete or the sources are secondary.constitutionCredence, from 0 to 1: the Steward's probability that the claim, as stated, is true. Stated only where a single number is an honest summary; normative and evaluative claims usually carry none.constitution
supportssupportssupportssupportscontradictscontradicts
Nothing in the graph builds on this claim yet.
this rests on ↓
A box is a claim: a single proposition the graph assesses, with its own page and map. Click any claim to centre the map on it.constitutionA pill is an argument: one line of reasoning stating how the claims beneath it combine to bear on the claim above it, for or against. Arguments are not destinations; click their claims to explore.constitutionThe claim traces to reliable primary sources through a clear chain of evidence.constitutionEvidence favors the claim, but the chain is incomplete or the sources are secondary.constitutionCredible evidence or argument exists on multiple sides.constitutionNo credible evidence found, though the claim is not contradicted.constitutionAvailable evidence weighs against the claim.constitutionInsufficient information to assess.constitutionNo current assessment. Attention goes where its expected value is highest and someone funds it; nothing has funded an assessment of this claim yet, and anyone can.constitutionverified factopen questionvalue premisethis provides evidence for the parentsteward instructionsthis argues against the parentsteward instructionsbackground the parent's framing takes as givensteward instructionsa load-bearing premise: the parent is false without itsteward instructionsFig. Detail falls off with distance; every claim is an address.