IT Home On October 13, according to reports by CCTV’s Chinese Voice today, the Chaoyang District People’s Court in Beijing recently concluded a labor dispute case involving the termination of a contract due to employees using AI for office assistance. In the first instance, it was determined that the insurance company’s unilateral termination was illegal. The court ordered the revocation of the termination notice, both parties to continue fulfilling their labor contracts, and ordered the company to pay the difference in year-end bonuses, performance wages, and wage losses. The judgment has taken effect. This case further clarifies the boundaries between compliant AI-based office work and illegal data leakage behavior.
According to the introduction, Ms. Wu worked as a legal officer at an insurance company, responsible for compliance audits and risk assessments. In July 2024, in order to optimize work processes and improve PPT production and data analysis efficiency, under the recommendation of company leadership, she uploaded the data tables of company audit issues that she had sorted and processed herself to a third-party AI platform to assist with office work.
One month later, the company issued her a “Notice of Termination of Labor Contract” on the grounds that Ms. Wu “seriously violated the rules and regulations” – as an employee in the Legal Compliance Department, she had signed a confidentiality agreement and participated in information security training, thus bearing confidentiality obligations higher than those of ordinary employees. However, she uploaded internal audit documents from the company to a third-party AI platform, which collects and retains user data for model iteration, posing a risk of information leakage.
The court held that Ms. Wu used a third-party AI tool recommended by company leadership, with the core purpose of assisting in completing her job, and there was no subjective intention to steal or disclose company information maliciously.
The evidence shows that although the insurance company has established a confidentiality system, information security management regulations, and signed confidentiality agreements with employees, none of the internal rules and regulations explicitly prohibit employees from using third-party AI tools to assist in their work. There are no special compliance rules or operating procedures for AI-based working scenarios, and no specialized training on the use of AI tools has been provided.
The court also noted that the company failed to provide evidence that the files uploaded by Ms. Wu are classified as legally or institutionally defined core confidential information, and it was unable to prove that such actions resulted in actual consequences such as information leakage, economic losses, or damage to reputation. Simply assuming a risk of data leakage based on the general rules for data retention on AI platforms is not sufficient to conclude that the employee’s behavior constitutes a “serious violation of regulations and rules.”
Presiding Judge Chai Xu said that the regulations of the defendant insurance company clearly consider “subjective intent” as a prerequisite for determining violations of the company’s rules and regulations. Therefore, the company has no right to unilaterally determine that an employee’s behavior is illegal afterwards.
The court ultimately determined that the insurance company’s unilateral termination of the labor contract was an illegal act. According to the law, the “Notice of Termination of Labor Contract” was revoked, and both parties continued to perform their labor contracts. The court also ordered the insurance company to pay Ms. Wu the difference in year-end bonuses, performance wages, and wage losses resulting from the illegal termination.
Assistant Judge Du Jie made suggestions regarding this case. For enterprises, in the face of the normalization trend of AI-based office work, they should abandon management models such as “one-size-fits-all prohibitions” and “retrospective accountability,” and actively improve the digital office compliance system. IT Home summarized it as follows:
Refine the list of confidential information, and clearly define the scope of core data that is prohibited from being entered or uploaded to the AI platform;
Establish guidelines for the use of AI tools and define boundaries for compliant usage;
Regularly conduct compliance training for digital office work, and clarify employees' responsibilities and risk boundaries.
For employees, they should actively understand the company’s digital office and confidentiality policies. It is strictly prohibited to upload corporate core secrets, undisclosed data, and confidential documents to public third-party AI platforms. At the same time, they should keep records of their office operations and conduct their duties in accordance with regulations.