Chen Meili spent seven years climbing the ranks of Hangzhou's new media industry. Starting from zero, she worked her way up to become the most reliable senior copywriter on her team — the one with all the viral hits, the steady hand, the institutional memory. She was the backbone of the content team.
Then the AI arrived.
At a company all-hands meeting at the beginning of last year, her boss announced they were building their own proprietary AI writing model. "AI will handle the grunt work," he told the room. "You'll all focus on creativity and deep work. Double your salaries, leave on time!"
Meili had her doubts, but she pushed them aside. Seven years of experience — the writing texture, the emotional calibration, the instinct for what goes viral — those were irreplaceable human skills. The machine could never replicate that.
Ironically, because she was the best writer on the team, her boss picked her to lead the AI training. "Key development role," he said. "Promotion and raise are first in line for you."
What followed was eight months of relentless, grinding work. The AI started as a total mess — producing hollow copy, logical errors, basic factual mistakes. Every piece it generated needed line-by-line corrections, logic restructuring, tone polishing, and style calibration before Meili could feed the improved version back into the system for it to learn.
Training a single AI piece took three times longer than just writing it herself. Meili worked 007 — zero days off, zero holidays, zero evenings leaving on time. Night after night, she was the last light on in the office building, eyes burning from screen blue light, typing correction after correction.
She poured everything she knew into it: the company's signature writing style, the content rhythm that hooked readers, the underlying logic of viral topics, the nuance of brand communication. Even her private, never-shared writing tricks — she broke them all down into steps and fed them to the system.
Her friends told her she was working too hard. She told them it would pay off when the AI was mature and she could finally take it easy.
After eight months, the AI was good. Really good. It had perfectly replicated Meili's writing style and mastered her viral formula. Brand posts, product copy, promotional articles — everything came out fluent, professional, zero errors.
Meili was proud. She joked with friends that she finally had a free assistant.
Then the layoffs came.
It happened the week before her performance review — the one where she was supposed to get her raise. HR called her in without warning, no lead-up, no cushion.
"Company restructuring," they said. "The copywriting position has excess capacity. Your role is eliminated. Don't come in tomorrow."
When she demanded to know why, the explanation was blunt and cold.
"The AI can now do all the copywriting independently," HR told her. "It works 24/7, needs no salary, never makes mistakes, and is multiple times more efficient. Since the machine can completely replace you, there's no reason to keep a human in the role. Put simply — you taught the AI everything it knows, so your value is now zero."
Meili wasn't the only one. Every senior employee who had participated in training the AI was systematically laid off. The most diligent, the most cooperative — the ones who had gotten the AI from zero to industry-leading — were the first to go.
"I used to stubbornly believe that creativity, texture, and warmth would always be human competitive advantages," Meili says. "But reality slapped me across the face. Top-tier creativity might not be replaceable — but the career experience of 99% of ordinary people? It crumbles in front of AI."
She spent seven years building her professional fortress. Then she spent eight months taking it apart, brick by brick, and feeding every piece to the machine. Now the machine sits in her seat, producing the work she used to make. And she's on the outside, wondering how so much hard work could lead to nothing.