In a troubling trend, tech employees in China are finding themselves in a position where they are being asked to train AI agents that could potentially replace them. This unsettling directive from management has led to a wave of introspection among workers, who had previously embraced AI technologies. A recent viral GitHub project named Colleague Skill, which humorously suggests that employees can distill the skills and personality traits of their colleagues into AI replicas, has amplified these concerns. Although initially intended as a parody, the project has struck a chord with workers who report being encouraged to document their workflows for automation purposes using AI tools like OpenClaw and Claude Code.
The Colleague Skill tool allows users to input the name of a coworker they wish to emulate, along with basic profile information. The system then imports chat histories and documents from popular Chinese workplace platforms like Lark and DingTalk, creating detailed manuals that describe the individual’s responsibilities and unique traits for AI replication. Tianyi Zhou, the engineer behind Colleague Skill, revealed that the project was inspired by a rise in AI-related layoffs and the trend of companies pushing employees to automate their own roles. While some users have found amusement in the concept, many are grappling with deeper issues surrounding job security, individuality, and dignity in the age of AI.
As AI agents become more integrated into workplace environments, tech workers are facing the reality that their roles may be supplanted by machines. Despite the capabilities of AI tools to manage tasks such as email responses and scheduling, many employees report that their effectiveness in professional settings remains limited. Hancheng Cao, an assistant professor at Emory University, notes that companies are encouraging the creation of work blueprints not just as a fad, but to gather valuable data on employee processes and decision-making. This can help organizations identify which tasks can be automated and which still require human input. Yet, the process of creating these AI agents often feels reductive to employees, reducing their complex roles to simple modules that can be easily replicated. The situation has spurred a mix of dark humor and resistance among workers, with some even developing countermeasures like the anti-distillation skill, aimed at sabotaging the process of creating AI replacements. As the conversation around AI and employment evolves, the implications for worker identity and dignity remain paramount, prompting calls for more robust discussions on these critical issues.
Source: Chinese tech workers are starting to train their AI doubles–and pushing back via MIT Technology Review