Anthropic report tracks AI use, hiring and work

Anthropic says junior employees’ share of technical-department employment fell from 21% to 14% since early 2025, while their share of technical hires fell from 27% to 9%.[1] Technical departments’ share of company employment also declined, from 61% to 56% over the last year.[1] Anthropic says the combined hours worked by autonomous and interactive agents first exceeded human work hours in March 2026.[1] By August 2026, it says employee-directed agents worked six times as many hours as their human handlers.[1]

Anthropic’s agent-hour calculation includes runtime, inference and active tool use, excludes waits longer than 30 minutes, and estimates human work using a 40-hour week.[1] NeoTeo reports that Anthropic estimated agent-management work made up 10% of interactive Claude sessions in fall 2025 and 22% in spring 2026.[2] Anthropic estimated that share by classifying a sample of 2,000 one-line chat summaries.[2] Anthropic also says code commits per engineer rose to five times their level six months earlier, with output increases in nontechnical work such as sales, legal and accounts-payable tasks.[1]

Anthropic compared 192 engineers who adopted coding agents with 74 who did not during the study period.[1] The company says adoption was not random: employees preparing for major projects may have been more likely to use the tools.[1] NeoTeo notes that this observational comparison does not isolate a causal effect of coding agents on output or hiring.[2] The report combines internal employment, AI-use, output, survey, communication, calendar and hiring records with 17 employee interviews.[1] Professionals should read the reported shifts as evidence to investigate, not as a settled forecast.

Anthropic says employees with at least a year at the company were 50% more likely to move across departments each month in 2026 than in 2025.[1] It says it has established an early-warning system to track jobs at risk of disruption and prioritized internal redeployments as jobs change.[1] Separately, a paper by Morgan Frank and coauthors reports that unemployment risk in U.S. occupations exposed to AI began rising in early 2022, and that graduate cohorts from 2021 onward entered those occupations at lower rates.[2] NeoTeo reports that Julián Posada questioned the report’s perspective and lack of independent outside assessment; Dylan Baker disputed the idea that human attention is simply a bottleneck in software work.[2]