In China, a project named Colleague Skill has gained significant traction as it trains AI agents with the skills and knowledge of employees prior to their departure from companies. The open-source development has sparked discussions on the social media platform RedNote, known locally as Xiaohongshu, and has garnered over 8,000 stars on GitHub.
The developer positions Colleague Skill as a means to facilitate a more effective farewell to colleagues. This service enables the transfer of an employee’s professional knowledge to an AI assistant, encompassing their information handling methods, communication style in chats and calls, and other job-related insights. To train the AI, users are required to upload various documents, including messages from work chats, spreadsheets, emails, audio recordings, and screenshots.
Colleague Skill is designed to write code based on technical specifications, answer inquiries, and even shift blame onto others. The project emerged amidst reports that an increasing number of employers were demanding employees document their work processes and decision-making logic systematically, often leading to their termination. This procedure was referred to by management as “process optimization,” but in reality, the data was utilized to train AI systems.
The reaction from users has been largely negative, with many labeling the situation as “worker distillation.” In response, a counter-project named Anti Distillation Skill has surfaced, aimed at protecting individuals from having their knowledge replicated. This tool allows users to rewrite archived documents to make them less beneficial for AI agents.
Additionally, some users have started sharing repositories that offer digital representations of former partners, bosses, or even themselves for business communications. Meanwhile, citizens in the United States express deep concerns over the advancements in artificial intelligence and the potential impact on unemployment.
