Anthropic staff expect far more AI automation than US workers—what does that prove?
A company survey compared 1,334 Anthropic employees with 10,980 US adults. The gap is large, but it records beliefs shaped by very different workplaces—not measured automation or a forecast of jobs lost.
The short version
3 key points
- What changed
- Morning Consult surveyed 10,980 US adults online in August and weighted the sample to national demographics; an unweighted internal poll reached 1,334 Anthropic employees.
- Why it matters
- For a worker deciding what to learn or whether a role is secure, the study shows that proximity to frontier AI changes expectations—but it cannot supply a personal redundancy probability or a timetable for retraining.
- What to watch
- Independent repeated surveys using identical questions, observed task and hiring data, occupational transitions, wages and re-employment duration—and whether expectations converge as workplace AI use spreads.
The disagreement is about the future, not today's capabilities
Anthropic employees and US adults broadly agreed on what AI can do now, then separated sharply on the pace of change. By 2030, 91% of staff expected AI to make a Nobel-level discovery, compared with 39% of US adults. Staff expected AI to be capable of performing 75% of their own work hours on average; employed public respondents expected 28%. Yet the result is a contrast between beliefs formed in very different workplaces. Anthropic staff use unusually capable internal systems and build the technology being judged. The survey did not observe AI completing those future tasks, count redundancies or measure productivity across the US economy.[1][2]
Who it affects
For a worker deciding what to learn or whether a role is secure, the study shows that proximity to frontier AI changes expectations—but it cannot supply a personal redundancy probability or a timetable for retraining.[1][2][3]
Keep in mind
Company-produced, non-peer-reviewed research; one US online public sample; an unweighted employee poll from an exceptional AI workplace; differing response formats and missing-response rules; expectations rather than measured future outcomes.
Multi-source analysis · 3 linked sources. Original source date: 16 October 2026.
Read the full report · 4 min The evidence, methods, implications and limitations
The disagreement is about the future, not today's capabilities
Anthropic employees and US adults broadly agreed on what AI can do now, then separated sharply on the pace of change. By 2030, 91% of staff expected AI to make a Nobel-level discovery, compared with 39% of US adults. Staff expected AI to be capable of performing 75% of their own work hours on average; employed public respondents expected 28%. Yet the result is a contrast between beliefs formed in very different workplaces. Anthropic staff use unusually capable internal systems and build the technology being judged. The survey did not observe AI completing those future tasks, count redundancies or measure productivity across the US economy.[1][2]
What the survey can tell workers and policymakers
Both groups linked greater expected automation of their own work with higher unemployment forecasts. US workers who thought AI would perform at least 80% of their hours forecast 12.2% unemployment in 2030 on average; those below 50% exposure forecast roughly 7%–8%. Anthropic staff showed a similar relationship. That does not mean either forecast will occur, but it identifies a practical concern: people who expect more of their tasks to become automatable also expect a harder labour market. Employers and governments should therefore track task change, entry-level hiring, retraining access and re-employment time with observed data rather than treating optimism about capability as evidence that workers will share the gains.[1][2]
The distinction that matters
Three layers that should not be collapsed into one forecast
The report connects survey answers to an economic model, but each layer supports a different kind of claim.
Belief
Ask what AI may do by 2030
Measures expectations in two unlike groups
Association
Compare expected task automation and unemployment
Shows linked views, not a causal effect
Scenario
Feed complete answers into a simplified model
Produces conditional growth paths, not forecasts
The headline growth figures come from a model, not economic data
Morning Consult collected 16,395 interviews from 11–23 August, removed repeat respondents and retained 10,980 people. The online opt-in sample was weighted by age, gender, race and ethnicity, education, region and income, but excludes people without internet access and cannot correct every selection difference. The 9–12 August staff poll was unweighted and offered different response formats; 12%–21% of public respondents chose ‘not sure’ on individual questions, an option staff did not receive. Only 3,259 US adults and 1,093 staff completed every input used for the economic scenarios. Those answers implied 5.4% versus 17.1% GDP growth in 2030 under Anthropic's model. The model's authors explicitly call these scenarios, not predictions. Confidence would rise through independent replication, identical question formats and comparison with observed workplace and labour-market outcomes over time.[1][2][3]
Watch the background, then check the evidence
Illustrative photo · not a video frame
UNESCO · 11 min 27 sec · Background explainer
UNESCO Talks: Interview with 2024 Nobel laureate Professor Daron Acemoglu
YouTube loads only when you choose; playback starts when you press play.
Economist Daron Acemoglu discusses how choices about AI affect workers, inequality and society. Pair the interview with our comparison of task exposure, hiring and skills evidence.
Evidence check: An expert interview from 2024, offering a perspective on policy and work rather than a measured forecast of job losses.
Published 2024-10-14. Source checked 14 October 2026. Check captions on the publisher’s video page.
Read the AI and jobs guidePhoto: Dylan Gillis / Unsplash · Licence
Evidence check
Company-produced, non-peer-reviewed research; one US online public sample; an unweighted employee poll from an exceptional AI workplace; differing response formats and missing-response rules; expectations rather than measured future outcomes.
Reporting basis: Multi-source analysis. Original source date: 16 October 2026.
Continue the story
Read next
Also about employment
What changed when Anthropic's agents worked six times human hours?
Junior technical hires fell from 27% to 9% while agent runtime surged. The internal study tracks a frontier laboratory—not the wider labour market—and runtime is not useful output.
Read the 4 min reportAlso about employment
If robots can do much physical work, why are so few tasks economical to automate?
Anthropic researchers rate 7,594 physical job tasks and estimate that robots can perform 74% in some setting, yet are currently cheaper than people for only 0.3% of all work. The gap between technical exposure and economic adoption is the result that matters for workers.
Read the 6 min reportExplore the full source and evidence record
What this means for people
- For a worker deciding what to learn or whether a role is secure, the study shows that proximity to frontier AI changes expectations—but it cannot supply a personal redundancy probability or a timetable for retraining.
Global context
The nationally weighted comparison represents US adults, not workers worldwide. Labour institutions, internet access, occupational mixes and exposure to frontier systems differ across countries.
What the evidence does not yet show
- Company-produced, non-peer-reviewed research; one US online public sample; an unweighted employee poll from an exceptional AI workplace; differing response formats and missing-response rules; expectations rather than measured future outcomes.
What to watch next
- Independent repeated surveys using identical questions, observed task and hiring data, occupational transitions, wages and re-employment duration—and whether expectations converge as workplace AI use spreads.
Living evidence record
Impact record IAI-0YYZJTF
Evidence stage
Observed
Confidence
Supported
Reporting basis
Multi-source analysis
Independent or research support
Not yet
Record status
Monitoring
Last checked
18 October 2026
Source trail
3 direct sources across 1 source type.
People impact
Documented in this record.
Uncertainty
Limits and next checks are explicit.
Stages describe the evidence available—not whether a technology is good or bad. See the public method.
Evidence trail
Sources used for this report
Links checked 18 October 2026
Research question, topics and regions
How expectations about AI capability, workplace automation and the economy differ between US adults and employees of a frontier AI laboratory
Topics: Work · Automation expectations · Employment · Surveys · Economic scenarios
Regions: United States · North America
This report is labelled multi-source analysis. We summarise and analyse source material in our own words; company statements remain attributed claims until independently supported. Translated summaries preserve the meaning of the original source and link back to it. Read our editorial standards.
The Impact Brief · Free
Keep following the stories that matter to you.
Get our weekday briefing on what changed and why it matters. Choose your topics below.
Reader commentary
What does this mean where you are?
Tell us what you have seen at work, in your community or in everyday life. A question is welcome too. No account is needed; we review notes before publishing them.
Explore commentary across the portal →Published reader notes
0No published reader notes yet. You can start the evidence-led discussion above.
Prefer a private correction or response? Contact the newsroom.