She Was Fired for Using an AI Tool at Work. A Court Gave Her the Job Back

A legal counsel at an insurance company in China has won her job back after being dismissed for using a third-party AI tool at work, in a case that turns on a simple question: can you break a rule that does not exist?

According to state broadcaster CCTV, the employee, surnamed Wu, uploaded a table of internal audit findings — which she had processed herself — to an AI platform in July 2024, to speed up slide-making and data analysis. The platform was not her own choice: her department head and her direct manager had recommended it more than once.

A month later the company dismissed her for “seriously violating information security rules”. She took the case to the labour arbitration commission in Beijing's Chaoyang district, and then to court.

The court found no evidence that she intended to leak anything, and held that this was not enough to establish the serious breach that would justify dismissal. It ordered the company to withdraw the termination, reinstate the contract, and pay about 46,000 yuan in lost wages and bonus differences. The company appealed; the second instance upheld the ruling, which is now final.

The reasoning is the useful part. The company had a confidentiality policy, information security rules and a non-disclosure agreement — but none of them prohibited using third-party AI tools, none addressed AI use specifically, and the company had run no training on it. It also failed to show that the files were core confidential material, or that anything had actually leaked, been lost or been damaged. Arguing that a platform's general data-retention policy creates a leak risk was not enough.

The wider lesson is not that uploading company data to an AI tool is safe. It is that “serious violation” has to point at a rule the employee was actually bound by — and that a company which recommends a tool without writing any rules about it leaves itself very little to stand on.