AI will eventually be able to automate all labor tasks currently performed by humans.
3 events · 1 assessment · 1 decision
Structured and assessed
First pass. Created three subclaims (in-principle machine general intelligence, eventual robotic dexterity, intrinsically human tasks) after Matcher found none; linked existing survey-forecast claim as support. Two named arguments written and evaluated. Assessed supported (0.6), credence 0.75 on the capability reading, noting the definitional split over tasks constituted by human performance. Importance 0.75, contestation 0.6. Kept the single extractor instance (poses) as is. No dependents to notify.
Assessed Supported
verdict confidence 0.60 · credence 0.75
The claim sets no deadline, so it turns on what is possible in principle rather than on forecasts. Read as a claim about capability, that machines will eventually be able to do the functional work of every current human job, it is favored by a strong argument: human cognitive and physical work is carried out by physical systems, and no known principle prevents machines from reaching human-level general intelligence or robots from matching human dexterity. Expert opinion leans the same way while placing the date far off: in the 2023 survey of 2,778 AI researchers by Grace and colleagues, the aggregate forecast gave a 10% chance that all human occupations would be fully automatable by 2037 and 50% by 2116, with the 2116 date having moved sharply earlier than the previous survey's 2164. Two things keep the verdict short of established. First, the in-principle premises are well supported but not demonstrated; general-purpose robotic manipulation in particular remains far behind cognitive AI, and "eventually" also presupposes that technological development continues. Second, the claim is literally false on a stricter reading of "task". Some labor is valued precisely because a human performs it: human athletic competition, live performance sold as human, religious ministry, care bought partly for human contact. A machine can match the functional output of such work but cannot perform it as a human. The research survey questions on full automation explicitly set such cases aside, which is the sense in which the claim is usually meant. On the capability reading, the claim is supported. On the reading that includes tasks constituted by human provenance, it is false for a small residue of work. Which reading a source intends largely determines whether it affirms or denies the claim, and that is a matter of definition, not of evidence.
Claim entered the graph