A satire project exposed a real workplace fear
A viral GitHub project in China has crystallized a growing anxiety inside the country’s tech sector: workers increasingly believe they are being asked to help build the systems that could replace them. The project, called
Colleague Skill
, presented itself as a way to clone a coworker’s skills and personality into a reusable AI agent. It was created as a spoof, but the reaction suggested the premise felt uncomfortably plausible.According to MIT Technology Review, workers told the publication that bosses are already encouraging them to document workflows so AI agent tools can automate specific tasks and processes. The joke landed because it echoed an emerging management pattern. In this version of workplace automation, employees are not simply adapting to new software. They are being asked to convert their own expertise into machine-readable instructions.
How the project works and why it resonated
The reported mechanics of Colleague Skill help explain why it spread so quickly. Users name a coworker, add profile details, import chat histories and files from workplace apps including Lark and DingTalk, and generate manuals describing not only duties but also quirks in communication style. The result is framed as a portable AI “coworker” able to help with tasks such as debugging code and responding instantly.
Even as a stunt, the project captured a real shift in how many companies now think about knowledge work. The emphasis is no longer only on general-purpose chatbots. It is on extracting tacit process knowledge from people and turning it into repeatable operational assets. That raises a different kind of concern from ordinary software adoption. The threat is not just efficiency pressure. It is the possibility that individual workers become templates.
One Shanghai tech worker, Amber Li, told the magazine she used the tool to recreate a former coworker as an experiment. She said the generated file captured the person’s habits surprisingly well, including how they reacted and even punctuation styles. That technical plausibility appears to be one reason the project became more than a joke.
Automation pressure meets labor insecurity
The creator of Colleague Skill, engineer Tianyi Zhou of the Shanghai Artificial Intelligence Laboratory, reportedly said the project was prompted by AI-related layoffs and by the tendency of companies to ask employees to automate themselves. That detail is important because it places the debate in a broader economic setting. AI adoption is not being received only as a productivity upgrade. For many workers, it is arriving in the context of job insecurity and management pressure.
The tension is especially sharp in the tech sector, where many employees are enthusiastic early adopters of AI tools. That makes the backlash more notable. Resistance is not coming only from workers hostile to automation or unfamiliar with the technology. It is coming from people who understand the systems well enough to see how quickly documentation, chat logs, and internal files can be turned into replacement infrastructure.
The article links this concern to the rapid popularity of AI agent tools such as OpenClaw or Claude Code. Once those tools became credible workflow engines, the act of writing down how a job gets done started to look less like training material and more like a migration path away from human discretion.
Dignity, individuality, and the future of office work
The debate that followed the project’s spread went beyond employment risk. It also touched on dignity and individuality. If a worker’s role can be distilled into manuals, examples, and stylistic markers, then the organization may start to treat professional identity as something modular and transferable. The worker is valued less as a person making judgments and more as a bundle of habits that can be captured, replayed, and scaled.
That does not mean AI agents are about to replace entire teams overnight. The source material does not support such a sweeping conclusion. But it does show that workers increasingly view knowledge capture as a contested process. Documentation used to be a sign of maturity and continuity. In an AI-heavy workplace, it may also be seen as a handoff mechanism.
The Chinese discussion is likely to resonate far beyond China. Many companies globally are exploring agentic tools that promise to encode best practices, standardize output, and reduce dependence on individual employees. The Chinese debate is an early indicator of how quickly those ambitions can collide with workforce trust.
What managers may be underestimating
The strongest lesson from the Colleague Skill episode is that organizations may be underestimating the social cost of automation programs that rely on worker self-extraction. Employees can often tolerate new tools when the goal is augmentation. They may react very differently when the process explicitly asks them to model themselves for substitution.
That difference may become one of the defining management questions of the agent era. The technical tools are improving fast. The harder problem may be convincing skilled workers that the systems they are training are meant to help them, not displace them. In China’s tech sector, that assurance already appears hard to sustain.
This article is based on reporting by MIT Technology Review. Read the original article.
Originally published on technologyreview.com
