AI Can Take Over 3.6 Billion Jobs In 15 Months, & Make Humans Unemployable For Ever: Anthropic Researcher

An Anthropic researcher has warned that rapid advances in artificial intelligence could eventually make almost every human permanently unemployable, leaving people dependent on government welfare for their livelihood.

Tao Lin, who works on AI safeguards at Anthropic, believes broadly superhuman artificial intelligence could emerge within approximately 1.3 years, potentially by the end of 2027. He argues that once AI systems become as capable as or better than the world’s top experts across almost every field, traditional human employment could face an unprecedented disruption.

Lin shared these views in an interview with Palisade Research, a US-based nonprofit that studies advanced AI capabilities, safety and potential risks.

AI Could Surpass Human Experts Across Almost Every Field

Lin’s prediction centres on the possibility of broadly superhuman AI. This refers to systems capable of matching or outperforming the best human experts across a wide range of professions, including programming, filmmaking and politics.

He believes the transition could initially involve workers losing their existing jobs and moving into different occupations. However, as AI capabilities expand, even those alternative jobs could eventually become automated.

In his assessment, the long-term outcome could be a world where human employment largely disappears and governments are required to provide financial support to people who can no longer find work.

Importantly, this is Lin’s personal forecast and should not be interpreted as an official Anthropic prediction or a confirmed timeline for AI development.

Why AI Will Not Replace All Jobs Overnight

Despite his aggressive timeline for superhuman AI, Lin does not expect widespread unemployment to happen immediately.

He distinguishes between AI’s technical ability to perform a task and its actual deployment across the economy.

For example, AI systems may become capable of performing complex intellectual work much sooner than companies can replace human workers in physical occupations.

Replacing people in manufacturing, construction, transportation and other physical industries would require the development, production and deployment of billions of advanced robots.

This infrastructure requirement could create a significant gap between AI becoming technically capable of replacing workers and that replacement happening at scale.

Anthropic’s Own Research Shows a More Measured Picture

Interestingly, Anthropic’s own research on employment presents a less dramatic picture of AI’s impact so far.

A study published in March by Anthropic economists Maxim Massenkoff and Peter McCrory found no systematic increase in unemployment among workers in occupations highly exposed to AI since late 2022.

However, the researchers identified early evidence that hiring of younger workers had slowed in occupations with greater AI exposure.

The study also highlighted the difference between what AI could theoretically accomplish and how businesses actually use it.

For instance, large language models could theoretically assist with 94% of tasks in computer and mathematics occupations, but actual workplace coverage was estimated at just 33%.

Researcher Wants AI Companies to Slow Down

Lin has also called on AI companies to significantly reduce the pace of their research and development.

He believes companies should slow their current AI research efforts to approximately one-quarter of their existing rate. This, according to him, would provide more time for safety testing and ensure that increasingly powerful systems remain under human control.

His warning adds to the growing debate over whether AI development is advancing faster than governments, businesses and society can prepare for its economic consequences.

Summary

Anthropic researcher Tao Lin predicts that broadly superhuman AI could emerge by the end of 2027, potentially making human workers permanently unemployable over time. He expects physical infrastructure limitations to delay widespread job replacement. While his forecast is highly speculative, Anthropic’s own research has identified early signs of slower hiring among younger workers in AI-exposed occupations.

Image Source


Leave a Comment