Safeguarding Digital Spaces

Article originally published April 16, 2025

DeepSeek, (Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd.) is a large language model Artificial Intelligence (AI) chatbot launched in February 2025. It has quickly become a popular AI application.

However, third party audits and UM internal investigations have identified numerous security and privacy concerns, including a lack of data transport layer security and several below standard data encryption practices. DeepSeek also collects extensive data (including personal information) and, as a result of insecure practices, this data is at risk.

These concerns have prompted the Canadian federal government and several Canadian universities to take precautionary action by banning the use of DeepSeek.

After careful review by IST and the Provost Office, several recommendations were proposed to safeguard the safety and privacy of UM members. These recommendations have been approved by the President’s Executive Team and the following regulations now apply to all UM students, faculty, and staff.

New regulations:

1. All members of UM must remove DeepSeek from any UM owned devices or any other devices (including personal devices) that connect to the UM network.
2. IST will block the DeepSeek domain and URL’s on the UM campus firewall.
3. IST will block DeepSeek (via Intune and Defender application firewall) on any devices that are managed with Intune or Defender applications.

Exceptions may be considered for academic purposes. Please contact Information Security and Compliance via the IST Service Desk for more information.

The UM Artificial Intelligence (AI) Committee is in the process of establishing an AI Framework for the university.

These restrictions are issued under the UM Use of Computer Facilities Policy and the Information Security Policy. Non-compliance may be addressed through appropriate remedies in accordance with UM policies.