A Muscovite with a salary of over 800 thousand was fired for leaking data to DeepSeek. The court sided with the company.
A court in Moscow ruled that the dismissal of a sales director at one of the engineering companies was lawful. The woman earned over 800,000 rubles a month and had worked at the organization for less than six months. The reason for her dismissal was the disclosure of official (commercial) secrets.
According to the case materials, the employee sent the company's confidential data in two directions at once: to the DeepSeek chatbot and to her personal email. After her dismissal, she attempted to challenge the order in court, demanding that the wording be changed to "by agreement of the parties" and that a "golden parachute" of 5 million rubles be paid. The court rejected all claims.
This case is one of the first public ones in Russia where dismissal is directly linked to the use of a foreign AI service for handling sensitive information.
Why this matters
Public neural networks like DeepSeek, ChatGPT, and similar tools are not designed by default for processing corporate secrets. Data that a user enters into a chat may:
- be stored on the servers of the developer company,
- be used for further training of models,
- theoretically become accessible during technical failures or privacy configuration errors (as has already happened with DeepSeek's public links).
Even if the service formally promises confidentiality, legally and technically control over the information passes to a third party. And in the case of DeepSeek, the servers are located under China's jurisdiction — this is an additional risk factor for Russian companies.
Employees should remember a simple rule: if information is not intended for public release, it must not be fed to a neural network. Not "for analysis," not "for nice formatting," not "to get the report done faster."
The court ruling shows that courts are willing to support employers in such disputes, especially when the facts of data transfer are confirmed. The high salary and short duration of employment at the company in this case did not work in the plaintiff's favor.
This is only the beginning. As generative AI grows in popularity, there will be more such cases. And the price of a mistake — for both the employee and the company — will only increase.
