Caiwei Chen reports in MIT Technology Review: Chinese tech workers push back against AI doubles

Caiwei Chen reports in MIT Technology Review: Chinese tech workers push back against AI doubles

On April 20, 2026, Caiwei Chen’s MIT Technology Review newsletter describes Chinese tech workers who are training AI doubles and pushing back against workplace automation. A GitHub project called Colleague Skill claimed to “distill” a worker’s skills and personality and replicate them with an AI agent; Chen says the project was a spoof, but it prompted soul-searching among early adopters. Tech workers told MIT Technology Review that bosses are encouraging them to document workflows for automation using tools like OpenClaw. Chen reports that many fear being “flattened into code” and losing their professional identity. Some are responding with tools intended to sabotage the automation process. The account links a spoof project to concerns workers say they already face at work: that documenting their expertise for automation could reduce their roles to something reproducible by software. It does not describe how the sabotage tools work, how widely they are being used, or what effects they have. The stakes the account names are workers’ jobs and professional identities; it does not establish that AI doubles have replaced workers. Text as retrieved on October 27, 2026; originally published April 20, 2026.

Publication: 2cd13ffb-07d5-3e95-53be-2acd41f9131c

attribution — uncertain: The newsletter is one piece: a roundup in the editors' own words, not a bare listing. The item's claims are fairly attributed to Caiwei Chen, whose byline appears under the blurb. However, the summary calls The Download 'Caiwei Chen's MIT Technology Review newsletter.' The page presents it as MIT Technology Review's weekday newsletter, and Chen is credited only for this item. Smallest repair: say 'MIT Technology Review's The Download newsletter, in an item by Caiwei Chen.'

This is today's edition of The Download, our weekday newsletter that provides a daily dose of what's going on in the world of technology.

summary — pass: The reading leads with the actor and the development. It explains that Colleague Skill was a spoof claiming to replicate a worker's skills and personality with an AI agent, a term defined on the page. It reports the claims that bosses are encouraging workflow documentation and that workers are pushing back. OpenClaw is identified in context as an automation tool. The question follows from gaps the source itself leaves open. The worker fear in 'flattened into code' is preserved.

Many now fear that they are being flattened into code and losing their professional identity.

date — pass: April 20, 2026 matches the UTC day of originalPublished, 2026-04-20T12:10Z. The source's 'Earlier this month' is not given false precision. The reading does not use later or undated passages such as the Deep Dive list. The retrieval date is disclosed. The page is a publisher page, not an archive copy. modifiedAt (11:44Z) precedes the publication instant, so no post-publication edit is indicated.

Earlier this month, a GitHub project called Colleague Skill struck a nerve

support — uncertain: The four claims match verbatim excerpts. However, the summary says 'The stakes the account names are workers' jobs and professional identities.' The source names professional identity and being 'flattened into code'; it does not name jobs. The interpretive line about roles becoming 'reproducible by software' is a reasonable paraphrase but is not labeled as interpretation. Smallest repair: drop 'jobs' or mark it as an inference.

Many now fear that they are being flattened into code and losing their professional identity.

media — pass: No media is selected and no verified candidates were supplied. This is appropriate.