Anthropic researcher warns AI could make humans permanently unemployable by 2027
Artificial intelligence could make humans permanently unemployable and leave billions dependent on government welfare, an Anthropic researcher has warned, predicting that broadly superhuman AI could emerge by the end of 2027.
Tao Lin, an AI researcher at Anthropic, made the prediction in an interview with Palisade Research, estimating that such technology could arrive within 1.3 years.
Lin defined broadly superhuman AI as systems capable of performing as well as or better than top human experts across almost every field, including programming, filmmaking and politics.
He warned that AI could eventually displace workers across most occupations, forcing people to repeatedly change jobs before becoming permanently unemployable and dependent on state support.
However, he said replacing workers in physical occupations would take longer, as it would require companies to develop and mass-produce advanced robots.
Lin also estimated a 20% chance that AI could "completely remove humanity from the equation". While he considered human extinction unlikely, he warned that AI could eventually control the world while humans continue living without meaningful power, comparing the scenario to the animated film WALL-E.
To reduce these risks, Lin urged AI companies to significantly slow research and development, suggesting they cut their current pace to one-quarter.
He argued that rapid model releases leave insufficient time for safety testing and that slower development would allow researchers to strengthen safeguards and maintain human control over AI systems.
Lin has worked at Anthropic for around 18 months, focusing on AI alignment, control research, and safeguards against harmful responses from its Claude models.