The Beijing Chaoyang District Court recently concluded a labor dispute, putting the boundaries of "using AI for work" under the spotlight. In July 2024, Ms. Wu, an employee at an insurance company, uploaded a spreadsheet of company audit issues that she had compiled herself to a third-party AI platform to help her make a PowerPoint presentation. Interestingly, her department and direct supervisor had repeatedly recommended such tools, believing they could significantly improve efficiency. But a month later, the company fired her for "seriously violating information security rules," and Wu promptly sued the company.
Court: If you didn't ban it, don't settle scores after the fact The company argued that the spreadsheet contained highly confidential core internal information and that uploading it risked leaking trade secrets. Wu countered that she used AI to complete her work, had her supervisor's recommendation, and that the file had been manually compiled, contained no sensitive content, and caused no actual loss. The court found that she had no subjective intent to leak secrets, so it was not enough to establish "serious misconduct" warranting termination. More importantly, although the company had confidentiality and information security policies, it had never explicitly prohibited the use of third-party AI tools, nor had it provided special training or clear guidelines. The evidence it submitted neither proved that the uploaded content was confidential nor demonstrated any actual loss.
A court ruling serves as a wake-up call for corporate management In the end, the court ordered the insurance company to continue performing the labor contract and pay wage losses, year-end bonus differences, and other amounts totaling more than 46,000 yuan. After the company appealed, the second-instance court upheld the original judgment. This case is like a mirror: now that AI has become an everyday office tool, if companies rely only on vague "information security" slogans but fail to write whether and how AI can be used into their rules and put that into training, when something actually goes wrong, they may end up paying for arbitrary dismissals.