· May 11, 2026
METR reports technical workers’ self-reported AI productivity gains
Markdown · JSON · Original source
published this survey on May 11, 2026. METR reports that technical workers surveyed in February–April described larger gains in work speed than work value: respondents’ median self-reported value change ranged from 1.4–2x across measures, while median speed change was 3x. METR distinguishes value—the value of work respondents believe they produce—from speed, or how long tasks take. It says speed measures are likely biased upward relative to value measures, partly because AI can make lower-value tasks quicker without raising the value of a person’s overall contribution. METR cautions that self-reports may not reflect actual productivity gains. Respondents may find questions difficult, and survey estimates have often exceeded field-experiment or quasi-experiment estimates, though the size of that difference is hard to determine. The convenience sample also plausibly suffers from significant selection bias. Still, METR argues surveys usefully complement other evidence because they are cheap, broad, and can address otherwise difficult questions. For organizations tracking AI’s potential to automate AI R&D, it recommends carefully defining what to measure and considering surveys of managers or productivity researchers.
Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims.
Questions using this source
These authored EM reads cite this source. Follow their assumptions and limitations; inclusion does not imply agreement.
Original source
· Article
Source details
Original source date: May 11, 2026 (2026-05-11T07:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-04T20:41:45.193Z. First observed: 2026-10-18T02:53:24.406Z. Retrieved: 2026-10-23T14:21:44.859Z.
Referenced by
Recorded citations and public file attachments. A reference alone does not establish agreement or influence.
No public references recorded here yet.
Limits
Text as retrieved on October 23, 2026; originally published May 11, 2026. The findings are self-reports from a convenience sample; says the results plausibly suffer from significant selection bias and may not reflect actual productivity gains. METR says it is difficult to determine how much surveys overestimate productivity gains relative to experimental data.
Open question
How closely would managers’ or productivity researchers’ estimates of AI’s effect on employees’ work value match workers’ self-reports?
Source-stated
METR reports that respondents’ median self-reported value change fell between 1.4x and 2x across its measures, while median self-reported speed change was 3x.
Participants self-reported a median 1.4–2x change in the value in their work due to AI tools. The median self-reported speed change (which we expect to be higher than value change) is 3x.
Source-stated
METR distinguishes the value respondents believe they produce from the time tasks take, and says speed measures are likely biased upward relative to value measures.
Speed measures are likely biased upwards with respect to value measures; however, value is also more opaque and harder for respondents to think about.
Source-stated
METR says the convenience sample plausibly suffers from significant selection bias.
Response rates are typically low – approximately 2% for respondents we email, although very significantly higher among METR staff and their networks – so our results plausibly suffer from significant selection bias.
Source-stated
METR argues that surveys are a useful complement to other evidence on AI capabilities, despite their limitations.
We think of surveys as a flawed but useful complement to other forms of evidence on AI capabilities.
Source-stated
METR recommends carefully defining the metric and considering surveys of managers or productivity researchers when tracking potential automation of AI R&D.
We particularly recommend carefully defining the metric to be estimated, such as by building on the survey questions discussed here. We also suggest prioritizing surveying managers or productivity researchers over individual contributors.
Source and contributor context
METR · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Counterfactual
An argument or scenario that considers what would have resulted if events had gone differently from how they actually did, such as if a policy had never been adopted; as an adjective, relating to this kind of thinking about what did not happen but could have.
METR
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Editorial provenance
Version 1. Publication ID: 2651171d-ed18-93a9-a6a3-5e7751e7993c.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-04T16:06:27.322Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_screen.
Observation: ae1ddb5e672301597bb07d2b342831234939440bace8a15e4778f180b9324e32. Source text hash: 8b47a32be31a689e627a34baf7b4d2e8998c64dc27e4c9442e0d431b9e502b8f.
Revision and source-check history
Version 1 · 2026-11-04T20:41:45.193Z
METR reports technical workers’ self-reported AI productivity gains
published this survey on May 11, 2026. METR reports that technical workers surveyed in February–April described larger gains in work speed than work value: respondents’ median self-reported value change ranged from 1.4–2x across measures, while median speed change was 3x. METR distinguishes value—the value of work respondents believe they produce—from speed, or how long tasks take. It says speed measures are likely biased upward relative to value measures, partly because AI can make lower-value tasks quicker without raising the value of a person’s overall contribution. METR cautions that self-reports may not reflect actual productivity gains. Respondents may find questions difficult, and survey estimates have often exceeded field-experiment or quasi-experiment estimates, though the size of that difference is hard to determine. The convenience sample also plausibly suffers from significant selection bias. Still, METR argues surveys usefully complement other evidence because they are cheap, broad, and can address otherwise difficult questions. For organizations tracking AI’s potential to automate AI R&D, it recommends carefully defining what to measure and considering surveys of managers or productivity researchers.
Questions using this source
These authored EM reads cite this source. Follow their assumptions and limitations; inclusion does not imply agreement.
Original source
· Article
Source details
Original source date: May 11, 2026 (2026-05-11T07:00:00.000Z).
Date basis: Date supplied by the publisher’s feed.
Added to Escape Meridian: 2026-11-04T20:41:45.193Z. First observed: 2026-10-18T02:53:24.406Z. Retrieved: 2026-10-23T14:21:44.859Z.
Limits
Text as retrieved on October 23, 2026; originally published May 11, 2026. The findings are self-reports from a convenience sample; says the results plausibly suffer from significant selection bias and may not reflect actual productivity gains. METR says it is difficult to determine how much surveys overestimate productivity gains relative to experimental data.
Open question
How closely would managers’ or productivity researchers’ estimates of AI’s effect on employees’ work value match workers’ self-reports?
Source-stated
METR reports that respondents’ median self-reported value change fell between 1.4x and 2x across its measures, while median self-reported speed change was 3x.
Participants self-reported a median 1.4–2x change in the value in their work due to AI tools. The median self-reported speed change (which we expect to be higher than value change) is 3x.
Source-stated
METR distinguishes the value respondents believe they produce from the time tasks take, and says speed measures are likely biased upward relative to value measures.
Speed measures are likely biased upwards with respect to value measures; however, value is also more opaque and harder for respondents to think about.
Source-stated
METR says the convenience sample plausibly suffers from significant selection bias.
Response rates are typically low – approximately 2% for respondents we email, although very significantly higher among METR staff and their networks – so our results plausibly suffer from significant selection bias.
Source-stated
METR argues that surveys are a useful complement to other evidence on AI capabilities, despite their limitations.
We think of surveys as a flawed but useful complement to other forms of evidence on AI capabilities.
Source-stated
METR recommends carefully defining the metric and considering surveys of managers or productivity researchers when tracking potential automation of AI R&D.
We particularly recommend carefully defining the metric to be estimated, such as by building on the survey questions discussed here. We also suggest prioritizing surveying managers or productivity researchers over individual contributors.
Source and contributor context
METR · Publisher
The publisher is identified here. No individual author has been inferred.
Source media
Media retrieval status: unchecked.
Terms and references
Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement.
Counterfactual
An argument or scenario that considers what would have resulted if events had gone differently from how they actually did, such as if a policy had never been adopted; as an adjective, relating to this kind of thinking about what did not happen but could have.
METR
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Editorial provenance
Version 1. Publication ID: 2651171d-ed18-93a9-a6a3-5e7751e7993c.
Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5.
Source check: passed by anthropic/claude-opus-5.5 on 2026-11-04T16:06:27.322Z. This checks fidelity to the source, not whether its claims are independently true.
Editorial selection: accept; reasons: bridge_screen.
Observation: ae1ddb5e672301597bb07d2b342831234939440bace8a15e4778f180b9324e32. Source text hash: 8b47a32be31a689e627a34baf7b4d2e8998c64dc27e4c9442e0d431b9e502b8f.
passed · 2026-11-04T16:06:27.322Z · anthropic/claude-opus-5.5
Publication: 2651171d-ed18-93a9-a6a3-5e7751e7993c
attribution — pass: The reading consistently attributes the survey, findings, caveats and recommendations to METR. The source is a single METR blog post with its own analysis, not a listing of other pieces.
In February–April 2026, we ran a survey of 349 technical workers
summary — pass: The reading leads with the actor (METR), the development (a survey of self-reported AI gains) and the central value-versus-speed distinction, explained in plain language. It preserves METR's caution and its argument that surveys remain useful. Counterfactual is defined on the page. It omits the 2025/2027 trajectory and the lower METR-staff answers, but those omissions do not distort the source. The question follows from METR's recommendation and is not leading.
We think of surveys as a flawed but useful complement to other forms of evidence on AI capabilities.
date — pass: originalPublished is 2026-05-11T07:00Z, which matches the URL path; the reading states May 11, 2026 at day precision. The source's March 2027 forecast is not used or presented as an outcome. There is no hindsight framing. The evidence kind is publisher-page, so no archive disclosure is required.
support — pass: Each claim is directly stated in the source: the median 1.4–2x value and 3x speed figures, the upward bias of speed measures, selection bias, surveys as a useful complement, and the recommendations on defining metrics and surveying managers. The caveat that surveys tend to exceed experimental estimates, with the size of the gap hard to determine, is faithfully reported.
However, it is difficult to determine the extent to which surveys overestimate productivity gains relative to experimental data.
media — pass: No media is selected, and no verified media candidates were supplied, so nothing is misidentified or irrelevant.
Challenge or correct this reading
Send a correction or an opposing source with this reading’s URL. Disagreement and uncertainty can remain unresolved.
METR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRAn argument or scenario that considers what would have resulted if events had gone differently from how they actually did, such as if a policy had never been adopted; as an adjective, relating to this kind of thinking about what did not happen but could have.
Definition referenceMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRAn argument or scenario that considers what would have resulted if events had gone differently from how they actually did, such as if a policy had never been adopted; as an adjective, relating to this kind of thinking about what did not happen but could have.
Definition referenceMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METRMETR studies what AI systems can do and how reliably they complete tasks, including work with human developers.
Explore METR