Job Displacement
4 gains7 Traces24 problems
Every measured gain and problem on the record for this topic, newest first. Updated .
One square per record, darker for stronger evidence:Evidence: 0 high17 moderate18 emerging
AI and job displacement
Hiring for entry-level and AI-exposed work looks weaker, and several companies tie layoffs to AI, but the broadest unemployment analysis found no major displacement so far. TRV-2026-1872TRV-2026-0740TRV-2026-1707 Most of the evidence is news reports, surveys and company statements; few records can separate AI from other pressures on jobs. TRV-2026-2215TRV-2026-0007
An IMF analysis of 40 million job postings in 19 Asian economies found demand in AI-exposed occupations fell nearly 10% relative to less-exposed ones, though postings measure recruitment rather than actual hiring. TRV-2026-1872 A study of 280 Chinese cities from 2011 to 2023 found AI pilot zones widened the pay gap between information technology and manufacturing workers, most in eastern and digitally advanced cities. TRV-2026-1520 A Higher Education Statistics Agency survey found the share of UK computer science graduates in professional coding jobs fell from about 40% to 28% last year; AI is suspected, not proven. TRV-2026-1061
Displaced workers in a 2,350-worker survey reported an average 26.7% income drop, and the authors describe an association with AI adoption rather than proven cause. TRV-2026-2027 Microsoft is reportedly offering buyouts to 60,000 employees to cut costs as AI use grows, yet a survey found 90% of 9,000 senior executives said AI had not yet raised productivity. TRV-2026-2282 At Lindy, which cut about a third of its staff and filled many roles with AI agents, workers say they now spend time checking the agents' output for errors. TRV-2026-2309 In a survey of 400 Pakistani university students, those more afraid of being replaced by AI reported more anxiety and lower life satisfaction. TRV-2026-1514
Interviews with 33 U.S. trucking stakeholders and case studies of six creative products rest on perceptions and small samples, and neither measured actual job outcomes. TRV-2026-1768TRV-2026-0350 A news report on young coders notes that no retraining program has yet shown it can reabsorb them at scale. TRV-2026-0007 A King's College London poll found one in three British university students expect AI job losses to trigger civil unrest, but it measures attitudes, not losses or unrest. TRV-2026-0073 Student surveys in China and Pakistan are cross-sectional, so they cannot show that fear of AI causes job anxiety. TRV-2026-2105TRV-2026-1514
A Stanford analysis found unemployment for the 20% most AI-exposed workers rose 0.77 points against 0.85 points for the least exposed, while a news report says employment for coders aged 22-25 fell roughly 20%. TRV-2026-0740TRV-2026-0007 Outside researchers dispute Anthropic's link between AI and its fall in junior hiring, and separate research suggests AI-exposed job markets weakened before ChatGPT's release. TRV-2026-2215 Centrica plans to cut 1,300 call centre jobs, 800 tied to AI tools, but its CEO says the cuts mainly reflect customers moving to digital channels. TRV-2026-0553 In a late-2025 survey of small EU firms, those using AI were more likely to report employment growth, an association the data cannot trace to AI. TRV-2026-1620
Job postings, graduate outcomes and company hiring data show less demand at the entry level, though causes are rarely separated from wider tech-hiring trends. TRV-2026-1872TRV-2026-1061TRV-2026-1707
See these 6 recordsEmployers from Microsoft to Centrica and Canva tie job cuts to AI, but several admit or leave open how much AI actually drove the decision. TRV-2026-2282TRV-2026-0553TRV-2026-2371
See these 8 recordsObservational studies and a survey report income losses and wider pay gaps, while an economist's model projects a widening gap between capital and labour income. TRV-2026-2027TRV-2026-1520TRV-2026-0378
See these 4 recordsA Trace is one record with both a gain and a problem, so it counts on both sides: 11 records show a gain and 31 show a problem. 23 context records are in the record list but not counted, because they report no measured gain or problem.
| Evidence | Gain side | Problem side |
|---|---|---|
| HighStrong study design, backed by at least two independent sources | 0 | 0 |
| ModerateA measured study or an official source | 5 | 16 |
| EmergingEarly, small, unreviewed or reported second-hand | 6 | 15 |
By type of evidence: 13 news reports; 8 surveys; 5 commentaries or narrative reviews; 4 investigations by reporters; 3 observational studies; 2 interview studies. How evidence is scored
Chartered accountants and audit-firm professionals who answered questionnaires said blockchain and AI improve audit quality, transparency and data reliability and lower operational costs.
In a late-2025 survey of small and medium firms in twelve EU countries, those using AI were more likely to report employment growth, though the data show an association, not proof AI caused it.
Anthropic says output per worker rose as it adopted Claude agents, including more code, more sales closed and more contracts processed, but the figures are its own and unaudited.
In a survey of 400 Pakistani university students, those more afraid of being replaced by AI reported more anxiety and career insecurity and lower life satisfaction, though the survey design cannot show the fear causes this.
Anthropic says its junior technical hiring fell from 27 per cent to 9 per cent between May 2025 and May 2026 as AI agents took on more work.
Surveys across several countries link fear of AI replacement to anxiety and lower career optimism, but they are cross-sectional and show association only. TRV-2026-1514TRV-2026-2105TRV-2026-1953
See these 9 recordsSome analyses find no major displacement so far, and a few record new solo businesses, rehiring of human experts or hiring growth among firms using AI. TRV-2026-0740TRV-2026-2423TRV-2026-1620
See these 4 recordsEach record sits in one thread at most. A record with both a gain and a problem counts on both sides. 2 records fit none of these threads.
Displaced workers in a 2,350-worker survey reported an average 26.7% income drop and 16.8-month job transitions, and AI adoption was associated with employment declines in manufacturing and administrative services.
1 record
Showing the records in one thread of the summary: Weaker hiring in AI-exposed and junior roles. Show every record
Anthropic published a report on 23 October 2026 describing how heavy use of its Claude agents has changed work inside the company. It says output per worker rose, AI agents logged more hours than humans from May 2026, and junior technical hiring shrank sharply.
The Guardian
ace