Colleague Skill: The AI That Worries Chinese Techs

AI Automation: A Reality Taking Hold in China

In China, technology workers are facing a new directive from their employers: to train artificial intelligence (AI) agents to automate their own tasks. This initiative, which may seem avant-garde, is actually provoking deep reflections among these professionals, who are typically enthusiastic about technological innovations. Recently, a GitHub project named Colleague Skill has garnered attention in China. This project aims to "distill" the skills and personality traits of colleagues to replicate them through an AI agent. Although originally conceived as a joke, this project has resonated strongly among tech workers, several of whom have told MIT Technology Review that their superiors are encouraging them to document their work processes in order to automate them using tools like OpenClaw or Claude Code.

How Does Colleague Skill Work?

The operation of Colleague Skill is relatively straightforward. The user begins by naming the colleague whose tasks they wish to replicate, then adds basic profile details. The tool then automatically imports the discussion history and files from popular work applications in China, such as Lark and DingTalk. From this data, it generates reusable manuals that describe the colleague's tasks and particularities, allowing an AI agent to replicate them.

A Project Born from a Joke, but Raising Serious Questions

Colleague Skill was created by Tianyi Zhou, an engineer at the Shanghai Artificial Intelligence Laboratory. Zhou explained to the Chinese media outlet Southern Metropolis Daily that the project was launched as a publicity stunt in response to layoffs related to AI and the growing trend of companies asking employees to automate themselves. Despite the initial humorous intent, the project has sparked serious debates about the dignity and individuality of workers in the age of AI.

Reactions from Tech Workers

Amber Li, a 27-year-old tech worker based in Shanghai, decided to experiment with Colleague Skill after discovering it on social media. She used the tool to recreate a former colleague, and within minutes, a file detailing that person's work methods was generated. Li was impressed by the tool's accuracy, which even captured her colleague's small quirks, such as their reactions and punctuation habits. However, she also felt a certain discomfort with the experience.

The Rise of AI Agents in China

The use of AI agents to replace colleagues could become a norm in China. Since OpenClaw gained popularity, employers are encouraging their employees to experiment with these agents. While these tools can perform various tasks, such as taking control of a computer, reading and summarizing news, responding to emails, or even making restaurant reservations, their utility in a professional context remains limited according to workers. The creation of detailed work manuals, as proposed by Colleague Skill, is an attempt to bridge this gap.

Companies and Automation: A Thoughtful Strategy

Hancheng Cao, an assistant professor at Emory University, believes that companies have valid reasons for pushing their employees to create detailed work plans. According to him, this not only allows for internal experience with the tools but also helps collect valuable data on employees' know-how, workflows, and decision-making patterns. This information aids companies in identifying which parts of the work can be standardized or codified, and which still require human judgment.

Workers Facing Alienation

For employees, creating agents or work plans can feel strange and alienating. A software engineer, who preferred to remain anonymous for fear of losing their job, shared their experience with MIT Technology Review. They trained an AI on their workflow and found the process reductive, as if their work was being simplified into easily replaceable modules. On social media, many workers express similar feelings through dark humor.

Countermeasures Emerging

In response to this trend, initiatives like that of Koki Xu are emerging. Xu, an AI product manager in Beijing, created an "anti-distillation" skill on GitHub to counter the reduction of workers to automatable skills. This tool, which can sabotage the workflow creation process for agents, has quickly gained popularity on social media.

Legal and Ethical Implications

Xu, who holds degrees in law, emphasizes that this trend also raises legal questions. While companies may claim ownership of work discussion histories and materials created on a work computer, capturing elements of personality, tone, and judgment complicates the issue of ownership. Xu hopes that Colleague Skill will encourage more discussions about protecting the dignity and identity of workers in the age of AI.

An Uncertain Future for Tech Workers

Amber Li, the tech worker in Shanghai, states that her company has not yet found a way to replace workers with AI tools, mainly due to their lack of reliability and the constant supervision they require. Although she does not feel immediately threatened, she senses a devaluation of her professional worth and is unsure how to cope with it.