university
IT helpdesk
737 5500 (5500)

For staff

At the University of Tartu, AI may be used in teaching, research, development and support work. The university’s position is that the use of AI must be purposeful, ethical and transparent, and its outputs must be approached critically.

For staff, this primarily means that AI can be useful, but it does not replace a responsible human. AI-generated responses may appear convincing but can, in fact, contain errors, outdated information or fabricated references. Therefore, staff must always assess whether the output is correct, relevant and suitable for use in university work.

The University of Tartu
  • encourages staff to use AI and ensures equal opportunities for doing so;
  • supports staff in using AI, incl. by offering general training opportunities and expert guidance on security-related matters;
  • provides the secure and stable infrastructure required for implementing AI and establishes governance processes and support services for its development and management;
  • develops guidelines and action plans for implementing AI and mitigating associated risks;
  • initiates discussions on the future of AI and its impact on Estonian society and culture, as well as on (Estonian-language) research and higher education.

AI can be used to support everyday work and to carry out routine tasks more efficiently.

For example, it is suitable for

  • drafting texts,
  • composing announcements and emails,
  • producing summaries,
  • gathering ideas,
  • simplifying complex information,
  • editing content, language revision and translation.


When using AI, substantive review must never be neglected. Good practice is to treat AI solely as an assistant; the final decision, verification, and responsibility always remain with the human.

The most important consideration is data protection. Sensitive or confidential data, such as passwords, health or social data, salary information, internal-use documents, technical configuration details, or copyrighted materials must not be entered into AI applications without proper authorisation.

The broader public framework in Estonia emphasises the same: when using AI, data protection requirements must be followed. The Ministry of Education and Research recommends using AI purposefully, ethically and critically, while the Data Protection Inspectorate highlights that the use of AI increases the need for strict protection of personal data and informed consent.

If a personal data breach occurs while using AI, it must be reported immediately to the IT helpdesk and the senior specialist for data protection.

What not to enter into an AI application and why?

Data type

Reason

Personal identification code

Data the disclosure of which may lead to identity theft

Health data, information about special needs or disabilities

Special categories of personal data the processing of which using AI is not permitted

Bank account and financial data

High-risk data the disclosure of which may cause financial harm

Information about disciplinary proceedings

Sensitive data the disclosure of which may cause harm

Grades, assessments and evaluation data of the teaching staff member or student

Data related to academic integrity and the protection of confidentiality

Agreements subject to confidentiality obligations

Data subject to intellectual property rights and contractual restrictions

Data concerning minors

Data requiring stricter protection and processing in controlled systems

Materials from protected research projects

Data subject to project-specific confidentiality requirements

Server configurations, API keys

Data that may create security risks and enable unauthorised access

Personal sensitive conversations

Data unrelated to work or study tasks that may pose risks to privacy

Data relating to court proceedings or offences

Special categories of personal data the processing of which using AI is not permitted

At the University of Tartu, the shared collaboration tool is Copilot Chat, an AI-based assistant. It is located within the Microsoft 365 environment and requires logging in with a University of Tartu account. The tool is free for the entire university community and provides a secure solution in which the entered data is not used to train models.

It allows users to

  • search for information,
  • summarise content,
  • generate text,
  • support ideation and text analysis,
  • simplify the completion of everyday work tasks.


Copilot Chat can be used on the desktop app, on mobile devices, and alongside other Microsoft 365 apps. For more information, see the IT helpdesk wiki.

Other AI-based tools we offer to our staff include

  • Microsoft 365 Copilot,
  • ChatGPT Education,
  • Google Gemini,
  • NotebookLM,
  • Azure OpenAI API,
  • GitHub Copilot,
  • SharePoint agents.

For more information and the ordering options, see the wiki page on artificial intelligence.

Internal training calendar

For UT staff

UT AI ambassadors info hub

Restricted access

AI-related news on the intranet

For UT staff
Chatbot Tiksu helps the university community
Since autumn 2024, an AI-based chatbot called Tiksu has been available on the University of Tartu IT helpdesk website. It has been trained to answer questions based on the guidelines available on the IT helpdesk wiki pages. In spring 2026, in collaboration between the Marketing and Communication Office and the IT Office, an intranet search assistant Tiksu was also launched. It generates responses based on intranet content pages and news articles up to two years old. The chatbot can respond in the language the question is asked in.

The text entered into chatbot conversations is not used to train the model.