Students always bear full responsibility for every thesis submitted, even in cases where they use generative AI in its preparation. Students are, at all times, obliged to carefully check all outputs involving generative AI, verify their correctness and incorporate generative AI into the thesis in a way that complies with the rules of academic ethics and the assignment of the thesis.

The author of the final thesis is, in all cases, the student, who is also responsible for the factual accuracy of all content and for the quality of the text. If errors caused by generative AI are present in the thesis, they cannot be invoked retrospectively; the student bears sole responsibility for the removal of such errors and for the final form of the submitted thesis.

In accordance with the Statement by Charles University, the use of generative AI tools in particular must be transparent.

This means that students should attach a brief statement at the beginning of their thesis in which they describe truthfully and in detail how, to what extent and for what purpose they used which generative AI tools during preparation of their thesis. (If necessary, a detailed version of this statement may be attached as a separate annex.)

This statement forms an important basis for assessing the originality of the thesis. Students can utilise the prepared template Declaration on the Use of AI in the Final Thesis, which helps them summarise all relevant information in a clear, structured manner.

How to write the Declaration on the Use of Generative Artificial Intelligence ?

Users must comply with the security rules when using generative AI tools. Some generative AI models are trained on the data that users enter into them, or their providers use the data entered by the user to “improve services”. They are not designed to ensure confidentiality or protect personal data; any information placed in a generative AI environment can be used to further train the model or can become public.

It is therefore prohibited to input any personal, sensitive or protected data into windely available generative AI tools, in particular information that is marked as protected in the Categorization of data table. Tools that are approved at the University, (e.g. MS Copilot when logging in with a university account - ukčo@cuni.cz) or locally operated generative AI models (provided that they are assessed as safe by CU) are exempt from this rule.

The student also undertakes not to use generative AI tools that have been evaluated by the Czech National Cyber and Information Security Agency (NÚKIB) and CU as potentially dangerous.

Caution and the responsible handling of data are basic prerequisites when working with generative AI tools on campus.

More abour data protection here.

If a student uses generative AI in the preparation of their thesis – for example, to generate images, diagrams or passages of text – they should give a proper description of such use of generative AI.

In such a case, generative AI should be cited as a tool in the Declaration on the Use of Generative AI.

In special cases when generative AI is used as a data source – for example, when researching the statistical properties of the generated text, the tool should be cited according to the citation standard chosen for the work (for example, according to APA, ČSN ISO 690 or other).

It is important to see generative AI primarily as a tool that modifies or generates content based on instructions and data that have been fed into it, and not as a reliable source of information. Only primary and secondary literature functions in this way.

Therefore, as when using any other digital or computational tool, students are also responsible for the resulting output. The output of generative AI represents the result of interaction between the author and the tool, and not a professionally verified or independent source of information.

It is important to verify whether the citation standard used contains specific recommendations for citing generative AI. Some standards describe this citation in detail, and others do not.

Further information about declaring AI tools here

Students are entirely within their rights to reject the use of generative AI in the preparation of their thesis without giving a reason. Some students always prefer their own independent work in all cases and may be concerned that the use of generative AI would negatively affect the development of their skills and critical thinking. Others may have reservations for moral, environmental or other reasons. All students are free to decide whether and to what extent to involve generative AI in the process of writing their final thesis, and this stance must be respected.

The areas of application for generative artificial intelligence in final theses are highly diverse, from linguistic editing of the text to data processing, through the summarisation of literature to assistance in formulating the structure or providing feedback on the text.

Considering this diversity, it is not possible to clearly and universally determine what is and what is not permissible in all fields and types of work.

It is therefore important that the student consults their supervisor about their specific intentions for using AI. Open communication with the thesis supervisor helps to avoid ambiguities or misunderstandings in the assessment and at the same time leads the students to responsible and thoughtful use of generative AI in an academic context.