"All Humans Will Become Unemployable": Anthropic Researcher's Warning

Virtually all humans could eventually become permanently unemployable, leaving governments to support people through welfare, according to Anthropic researcher, Tao Lin, who works on AI safeguards at the company behind Claude.
Lin made a stark prediction about the future of human employment, arguing that AI could eventually outperform the world's best human experts across almost every field.
"Pretty much all humans on Earth will become permanently unemployable and will need to strictly rely on government welfare to sustain themselves. Human employment will go away," Lin told Palisade Research, a US nonprofit that studies dangerous capabilities, cyber risks, and controllability in frontier AI systems.
Watch our quick take:
Lin also believes the world could be just 15 months away from broadly superhuman AI - which is basically AI that can match or outperform the world's best human experts across almost every field, including highly specialised professions.
To be sure, this is a personal forecast from a researcher working at one of the world's leading AI companies, not an official Anthropic prediction.
Lin does not believe that the arrival of superhuman AI will immediately make human employment obsolete.
He expects a considerable gap between AI acquiring the ability to perform most human jobs and that capability translating into widespread unemployment because there is a gap between capability and real-world deployment, partly because physical infrastructure and robots are not yet scaled up enough
Anthropic's Take On AI And Jobs
Anthropic's own research on AI and employment offers a considerably more measured picture of the disruption so far.
In a study published in March, Anthropic economists Maxim Massenkoff and Peter McCrory examined how AI was being used across different occupations, comparing its theoretical capabilities with actual workplace usage.
The researchers found no systematic increase in unemployment among workers in highly AI-exposed occupations since late 2022. They did, however, find suggestive evidence that hiring of younger workers had slowed in occupations more exposed to AI.
Anthropic also found a significant gap between what AI could theoretically do and what it was actually being used for.
For instance, although large language models (LLMs) could theoretically assist with 94% of tasks in computer and mathematics occupations, Anthropic's measure of actual workplace coverage put the figure at just 33%.
The study also found that programmers were among the most AI-exposed occupations, with 75 per cent task coverage.
In other words, AI may be getting better at performing tasks that humans are paid to do, but that does not automatically translate into entire occupations disappearing overnight.
A separate study by the Forecasting Research Institute, also published in March and revised in May, surveyed academic economists, AI experts, policy researchers and superforecasters about AI's potential economic impact.
The researchers examined a hypothetical scenario in which AI systems surpass human performance across many cognitive and physical tasks.
Under this rapid-progress scenario, economists projected that the US labour force participation rate could fall from around 62 per cent to 55 per cent by 2050. Roughly half of that decline, equivalent to around 10 million jobs, was attributed to AI.
The study also found that experts generally favoured targeted measures such as worker retraining, while the general public was more supportive of broader interventions, including universal basic income and job guarantees.
KioskNews shows a cleaned-up reading view extracted from the publisher’s page — the original always lives on their site, not ours.