AI-powered employee replicas challenge existing labour norms
According to a Sixth Tone report, Chinese companies are increasingly using AI systems to preserve employee knowledge by creating digital profiles of departing workers. The practice aims to improve continuity and productivity but is also raising concerns about automation, accountability, and the future role of human expertise in the workplace.
Chinese companies are increasingly using AI systems to capture and replicate employees’ knowledge before they leave, raising new questions about workplace automation, professional identity, and accountability.
According to a report by Sixth Tone, some workers at major technology companies are being asked to build AI versions of themselves as part of the handover process. These digital profiles are trained on code, documents, research materials, and internal communications to replicate how employees solve problems and collaborate with colleagues.
One former algorithm engineer at Baidu said colleagues used his work history to build an internal AI agent that could handle much of his previous workload. After his departure, coworkers could continue interacting with the AI system to ask technical questions, assign tasks, and access knowledge accumulated during his time at the company.
The practice reflects a broader trend in China’s technology sector, where organisations are increasingly building internal AI tools designed to preserve institutional knowledge and improve productivity. Chinese social media users have dubbed the phenomenon zhengliu, or ‘distillation’, borrowing an AI term for transferring knowledge from a larger model to a smaller one.
Interest in the concept grew after an engineer at the Shanghai AI Laboratory created a satirical project called colleague.skill, which proposed turning former employees into AI agents trained on their files, chat histories, and work habits. The project attracted significant attention online and inspired similar experiments involving AI representations of colleagues, public figures, and other individuals.
While proponents argue that such systems help preserve expertise and reduce disruption when employees leave, some workers have expressed concerns that they may be helping train their own replacements. According to the report, some employees have begun withholding parts of their workflows from AI systems, while others have experimented with tools designed to obscure how they operate.
The growing use of employee-based AI systems has also raised legal and ethical questions. Experts cited by Sixth Tone noted that existing legal frameworks may not clearly address issues related to AI models trained on an employee’s behaviour, decision-making processes, and accumulated professional experience.
Questions remain about accountability if AI-generated work results in errors, data leaks, or intellectual property disputes. Concerns have also been raised about the impact on workers’ sense of professional identity and ownership over their expertise.
Why does it matter?
The development comes as companies worldwide explore new uses of generative AI to automate tasks, preserve institutional knowledge, and improve efficiency. However, the emergence of AI systems designed to replicate individual employees highlights a growing debate over the relationship between human expertise and increasingly capable workplace AI tools.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our chatbot!