For years, a buzzy Silicon Valley startup called Mercor has been hiring an army of desperate job-seekers — often including educated and underemployed experts — to train AI models to replace them in the workforce.

It’s a grim facet of an AI-dominated future in which the business world continues to push for automation, hoping to wean itself off relying on pesky and expensive human labor once and for all.

As MIT Tech Review reports, an eerily similar situation is now playing out in China. Workers told the publication that their bosses are directing them to painstakingly document their workflows with the eventual goal of automating specific tasks using AI agents, such as OpenClaw, an open-source piece of software that has become immensely popular in the country.

Chinese employees already got an early glimpse of what an AI agent-led future could look like. A GitHub project called Colleague Skill, which was originally set up as a joke, went viral on Chinese social media. It works by ingesting the chat history and profile details of a specific coworker, then automatically spitting out workplace manuals that describe their tasks in stunning detail.

Purportedly AI-related layoffs roiling the tech industry reportedly inspired the tool’s creator, Tianyi Zhou. But while it was meant to poke fun at the trend, the tool also sparked a fierce debate over the future of human agency — and dignity.

“It is surprisingly good,” Shanghai-based tech worker Amber Li told MIT Tech of the software. “It even captures the person’s little quirks, like how they react and their punctuation habits.”

It’s a pertinent topic in the country considering the recent frenzy surrounding OpenClaw. The spread of countless agents grew so quickly this year that government agencies and state-owned enterprises started warning their staff not to install the software on their devices, citing cybersecurity risks including leaks and the mistaken deletion of data.

While businesses are incentivized to continue pushing for automation, streamlining workflows, and standardizing systems, employees are unsurprisingly far less receptive.

Some have even started to build tools to sabotage the creation of AI agents to replace human workers, according to MIT Tech.

“I originally wanted to write an op-ed, but decided it would be more useful to make something that pushes back against it,” AI product manager Koki Xu, who created a tool that rewrites worker manuals into non-actionable language, told the outlet.

While researchers continue to debate how effective AI agents will be in actually replacing human workers wholesale, employees are pushing to be part of the discussion, regardless.

“I believe it’s important to keep up with these trends so we (employees) can participate in shaping how they are used,” Xu added.

More on automation: Companies Just Learned a Brutal Lesson About Training AI to Do Human Jobs