Anthropic report puts workplace automation on a one-year clock

145

TL;DR:

Over 35% of Claude users expect AI to handle most of their work within a year, according to Anthropic, signalling a shift from AI as a productivity assistant to AI as a task owner. The biggest impact may fall on job design, training, and early-career roles.

Article:

More than 35% of Claude users expect artificial intelligence to perform most of their work tasks within the next year, according to Anthropic’s latest Economic Index report. The finding matters because it moves the AI-at-work debate from distant disruption to near-term workflow redesign.

The report, based on about 9,700 Claude users linked with privacy-preserving usage data, found that nearly six in 10 respondents expect AI to handle a larger share of their work over the next 12 months. Anthropic noted that “over 35% predicted that AI would be able to do most of their work.”

For employers, the signal is clear: AI adoption is no longer just about productivity tools. It is about deciding which tasks should be automated, which require human judgment, and how roles must change before disruption arrives through attrition, outsourcing or silent deskilling.

The report also complicates the fear narrative. Anthropic found that users who delegate more tasks to Claude were often more optimistic about pay, job security and the value of their skills. But early-career workers reported higher concern about job loss, making entry-level white-collar roles the pressure point to watch.

The takeaway is not that AI will replace entire jobs overnight. It is that AI workplace automation is already eating into task bundles. Companies that redesign jobs, training and accountability now will capture productivity gains. Those that wait may find their org charts changed by default.