Chinese Workers Create Tool to Block AI Job Replacement

When Koki Xu's boss suggested employees document their workflows for AI agents, she didn't get angry. She got creative.

The 26-year-old AI product manager in Beijing spent an hour building an "anti-distillation" tool that sabotages attempts to replace workers with AI. Her video about the project drew over 5 million likes across Chinese social media platforms.

Xu's tool came in response to a growing trend in China's tech sector. Companies have been pushing employees to use tools like Colleague Skill, which scans workplace chat histories and files to create detailed manuals of how workers do their jobs, including their unique quirks and communication styles.

The idea started as satire. Engineer Tianyi Zhou created Colleague Skill as a spoof, prompted by AI-related layoffs and companies asking employees to automate themselves out of jobs.

But the joke touched a nerve. Amber Li, a 27-year-old tech worker in Shanghai, tested the tool on a former colleague and found it surprisingly accurate, capturing everything from technical processes to punctuation habits.

Why This Inspires

What makes this story hopeful isn't just the pushback. It's the thoughtfulness behind it.

Xu, who holds degrees in law, sees deeper questions about worker dignity and identity. While companies might claim work materials as corporate property, she points out that capturing someone's tone, judgment, and personality raises murky legal and ethical questions.

Rather than writing an angry op-ed, Xu built a practical solution. Her anti-distillation tool lets users choose light, medium, or heavy sabotage modes depending on their situation. It rewrites workflow documentation into generic, non-actionable language that produces useless AI replacements.

The response shows workers aren't rejecting AI outright. Xu herself uses seven AI agents across her devices. Li finds AI helpful for debugging code.

They're rejecting the dehumanizing parts. One anonymous software engineer told MIT Technology Review that training an AI on their workflow felt reductive, as if their work had been "flattened into modules" designed to replace them.

Hancheng Cao, an assistant professor at Emory University studying AI and work, notes that while companies gain valuable data on workflows, the human cost matters too. Workers are using humor and creativity to assert their value.

"I believe it's important to keep up with these trends so we can participate in shaping how they are used," Xu says. Her tool proves that workers aren't just subjects of automation. They're active participants in deciding how technology reshapes work.