Nearly a year after leaving her job, video editor Li Dao discovered her voice was still appearing in internal meetings at her former company. She realized the company had used AI to replicate her voice and integrate it into new products.
'I only found out when others told me', she said. Since then, every evening Li often browses short promotional videos to find evidence of her voice being used without permission.
Today, AI is not only replacing repetitive tasks but is also replicating human experience, skills, and ways of thinking. Companies are asking employees to convert their intuitive abilities and problem-solving skills into commands for AI to learn. The goal is to retain employees' skill sets within the system even after they have left.
Xu Kejia, a Stanford University student who interned at a technology conglomerate, stated her task was to teach AI to write like a human. 'I can write faster myself, but the company needs someone to translate creativity into processes for AI', Xu said. She believes that the first people to be laid off will be those who trained the system.
In the US, the expertise of a group of security engineers in Seattle was encoded into AI skills to automate source code checks. In 4/2026, Meta collected employee keystrokes and mouse clicks to train AI to 'learn how smart people work', sparking internal protests.
 |
An employee demonstrates the operation of a humanoid robot at the China Import and Export Fair, 4/2026. *Thinkchina* |
Not only employees, but leaders are also being replicated. A Gartner report from early 2026 confirmed that 'digital twins' are being developed to recreate the management styles of outstanding CEOs. For example, when renowned admissions consultant Zhang Xuefeng passed away, the Zhang Xuefeng.skill project immediately emerged, extracting data from books and interviews to create an AI consultant in his distinct style.
Illustrating this trend is the colleague.skill tool developed by engineer Zhou Tianyi. The system absorbs employees' 'digital footprints' from chat logs to documents, simulating work habits to automatically write reports. While its initial purpose was to preserve collective knowledge, this tool has raised numerous ethical concerns.
In response, workers have begun creating counter-tools. Programmer Deng Xiaoxian released anti-distillation.skill, an application that helps employees replace core knowledge with 'correct but meaningless' content before submitting AI training data to companies. The tool garnered over 4 million views on GitHub in four days. Researcher Lu Cheng also released the keep-a-hand.skill tool to help individuals retain personal skills.
Lawyer Yu Zehui stated that civil lawsuits often drag on because detecting AI infringement requires manual effort. Expert Wang Tianyu from the Chinese Academy of Social Sciences proposed a 'data dividend' mechanism, demanding compensation for workers when AI replicates their skills.
Sociology professor Sun Liping of Tsinghua University, China, believes AI should only perform tasks humans do not want to do, rather than replacing their jobs. Researcher Lu Shiyu stated that the movement against 'skill distillation' serves as a reminder to businesses that workers are not machines.
Bao Nhien (According to Caixin)
