AI Could Leave Skilled Workers Permanently Jobless

An Anthropic safety researcher projects broadly superhuman AI by 2027, a research-stage forecast rather than a measured or deployed capability.

Updated on Nov. 22, 2026 in Artificial Intelligence

AI Could Leave Skilled Workers Permanently Jobless

Live Poll

Are you worried AI could make your own job obsolete within the next few years?

Anthropic researcher Tao Lin warned that humans could become permanently unemployable as AI systems surpass skilled professionals, in an interview with AI safety nonprofit Palisade Research released earlier this month. The views are the stated opinions of one researcher and are not a shipped product or an established forecast.

Why it matters

Lin's warning reflects a view held inside at least one leading AI lab that capability gains could outpace the labor market's ability to adjust. The projections are forecasts from a researcher whose work spans automated alignment research, AI control, and safeguards, so they signal a safety-community trajectory rather than a measured economic result.

The evidence is an individual researcher's projections in an interview, not a measured capability or benchmark, so no performance figures are supplied. Lin's specific estimates (a roughly 20% chance of loss of meaningful human control and broadly superhuman AI by the end of 2027) remain unverified, and the interview does not report the data behind them.

The players

Anthropic

An AI company that develops the Claude AI assistant and conducts alignment, control, and safeguards research.

Palisade Research

An AI safety nonprofit that interviewed Anthropic researcher Tao Lin on AI risk and labor-market effects.

The details

Lin's current work covers automated alignment research, AI control (monitoring deployed AI models to identify potentially harmful behavior), and safeguards (examining requests made to Anthropic's AI assistant, Claude). Lin said that physical-labor jobs would be phased out more slowly because they depend on advances in robotics and manufacturing infrastructure, while cognitive work tied to software could be displaced sooner. He also said governments could be required to financially support populations that can no longer sustain themselves.

Timeline

  1. Earlier this month (November 2026): Palisade Research releases its interview with Tao Lin.

  2. End of 2027: Lin's projected arrival date for broadly superhuman AI.

The Tech Race

Anthropic's automated alignment, AI control, and safeguards research places Lin's views within a safety-focused strand of the frontier AI race rather than a product competition. The projections therefore reflect how safety researchers at one lab are framing the timeline against which their control work is measured.

The projections describe no product or policy that reaches users now, and no timeline for labor-market effects has been set. Readers should treat the claims as one researcher's forecast about future employment rather than a near-term change in available AI tools.

The takeaway

Lin's forecasts put broadly superhuman AI by the end of 2027, a date readers can check against released model capabilities and any published evaluations. Watch for concrete labor-market data and for whether Anthropic's control and safeguards work publishes results bearing on the projected timeline.

Further reading

For broader context on where these systems are heading, see the Artificial Intelligence section.

Source note: This article includes information reported by TimesNow.

Live Poll

Are you worried AI could make your own job obsolete within the next few years?