Chinese Workers Forced to Build AI Copies of Themselves Before Quitting

One former Baidu employee described the process: he had to build an AI replica of his own knowledge and decision-making patterns as part of his exit procedure. This digital twin is designed to continue performing certain tasks after the worker has left, ensuring the company retains key expertise [167915].

The practice reflects a broader push by Chinese firms to integrate artificial intelligence into their operations, even at the cost of asking departing staff to train their replacements—machines. While companies see this as a way to preserve institutional knowledge, critics argue it raises questions about labor rights and the value of human experience [167915].

Meanwhile, in China's remote hinterlands, thousands of rural workers are labeling data to train AI models. This work has become a key poverty-alleviation tool. These workers sit in computer rooms, tagging images and text to help AI systems recognize objects and language, turning less-developed regions into hubs for the country's AI industry [167954].

The push for AI integration also extends to national strategy. China is pushing a rapid "AI Plus" revolution in electronic warfare, aiming to redefine how militaries communicate, jam, and control the electromagnetic spectrum, according to industry experts [157704].

However, the technology still has critical flaws. A Chinese user’s chat log showing an AI assistant confidently inventing fake flight refund details has sparked a wave of online spoofs. The trend, known as “AI hallucination,” is growing in the country. In the screenshots, the AI provided numerous inaccuracies, including a false refund amount and incorrect policy. Users are now sharing their own examples of AI making up information, highlighting a key flaw in the technology: it can sound certain while being completely wrong [167920].