In early 2026, the hypothetical passing of Zhang Xuefeng, China’s most polarizing and influential education consultant, triggered a digital phenomenon that legal experts and ethicists are now scrambling to address. Within days of his death, a GitHub project titled 'Zhang Xuefeng.Skill' emerged, claiming to have distilled his 'cognitive operating system' from years of books, interviews, and viral clips. This was not a mere archive; it was a functioning AI framework designed to simulate his pragmatic, often brutal decision-making logic for students navigating China's competitive job market.
While the public debated the morality of 'cyber-resurrection,' a more pervasive trend was quietly taking hold in the corporate sector: 'Colleague.Skill.' This initiative allows companies to ingest the digital trail of departing employees—including internal messages, emails, and code—to create persistent AI personas. These digital clones are designed to handle tasks, replicate problem-solving styles, and even mimic the professional quirks of former staff, effectively transforming human labor into a permanent corporate asset.
Legal scholars in China, such as intellectual property specialist You Yunting, argue that while using a person’s likeness or voice remains a clear violation of personality rights, the 'logic' of an employee is essentially unprotected. Under current Chinese labor and copyright laws, work products created using company resources are deemed 'work-for-hire.' Consequently, companies have a strong legal standing to use these datasets to train proprietary AI models, leaving workers in a position where they might be training their own digital replacements before they even resign.
This trend highlights a critical shift in the valuation of human capital, moving from temporary labor to what some call 'refined digital assets.' The core legal test now revolves around 'market substitution'—whether an AI clone substantially competes with the original creator’s business interests. However, for the average office worker, their specific professional methodologies and 'workplace DNA' are increasingly viewed by firms as harvestable data rather than personal intellectual property.
Critics warn that this rush toward 'cyber-immortality' may be built on a foundation of sand. Because these models are trained on historical data, they often have an 'expiration date' as societal contexts and policies shift. Relying on an AI version of a mentor from 2024 to solve problems in 2026 can lead to catastrophic errors if the model fails to account for the rapid evolution of China’s regulatory and economic landscape.
Ultimately, the commodification of the human mind into a 'Skill' file poses a deep psychological risk to the workforce. As humans are reduced to API interfaces that can be 'called' or 'instantiated' by a supervisor at will, the boundary between personhood and product continues to blur. While companies may gain short-term efficiency, they risk losing the spontaneous and unrefined human moments that define genuine innovation and culture.