Unless otherwise stated by your Study Board, you are permitted to use generative AI services in your project work for e.g. generating ideas or information searching. In this connection, it is important that you state your sources correctly (see paragraph on "Good academic practice regarding generative AI").

Pay attention to the extent to which you can use generated text directly, or paraphrased, in a project report/paper without risking plagiarism according to the disciplinary rules.

Your Study Board may have stipulated further guidelines regarding use of generative AI in project work and at exams. If that is the case, you should follow the guidelines stipulated by your Study Board.

Openness and transparency is essential not only in terms of complying with formal rules and guidelines, but also in terms of group work and task solution. It is important that you and your project group are in agreement when stating the project group's use of generative AI. This will reduce the risk of misunderstandings and conflict.

Generative AI is available at many different platforms, but it is not always clearly stated and can be difficult to tell. You can find the technology :

1. In services clearly defined as generative AI services e.g. ChatGPT, Microsoft Copilot, MidJourney, DALL.E and Google Gemini
2. As an integrated element in applications, browser plug-ins, services or software such as Microsoft Bing that displays AI generated elements and e.g. search results side by side. This also includes services that automatically generate summaries of scientific material.
3. Embedded in applications, services or software as a function to generate text, images or code like "type ahead" or suggestions like GitHub CoPilot

Pay attention to permitted exam aids for e.g. information searching. You may not be permitted to use a service that combines AI generated search results with regular search results.

Pay close attention to point 3 regarding the software you use for written exam papers. You may have to switch off certain functionalities in the software to make sure that you do not break the rules by mistake, or produce text that falls into the category plagiarism! 

Finally, in some cases a student may be covered by specific exam conditions due to an individual evaluation and exemption that allows certain exam aids although they use generative AI.

Whether or not you are permitted to use generative AI or similar technologies as an aid for text revision is stated in the module description of your curriculum. If use of generative AI is permitted, you must declare openly how you have used the technology and for what purpose.

Unless it is clearly stated as a permitted exam aid in the module description, you are not permitted to apply methods based on generative AI to revise, interpret, or create source code during an exam.

Take notice that code published under open source licenses may constitute part of the output from e.g. GitHub Copilot, Codeium or similar services. It you reuse large portions of AI generated code, it is recommended that you investigate the code to determine whether or not the original code library (and the corresponding license) can be included in your work.

Whenever you use generative AI services based on large language models to generate answers, it is important to pay attention to the validity of the answers. Consider the following:

You do not know the source

You are probably not familiar with the scope of the material that the service uses as its source. The material is often based on texts from news media, publicly available research, social media, debate fora and similar. Usually there is a fixed date for the model's last training session.

AI services hallucinate

AI-models can hallucinate and fabricate sources or facts! Therefore, it is important that you use other tools or known knowledge to verify statements and sources, and to make references to the original sources rather than to the AI generated answer. Furthermore, it is important that you declare any use of generative AI in your description of methodical approach.

Ethics

The service you use has a predefined code for assessing what is right and what is wrong. This affects the answers in such a way that original views from the sources may be suppressed or omitted in the AI generated answer. As such there are ethics to consider in terms of e.g. including different points of view.

Method

If you use generative AI to e.g. create an overview of a scientific field, and the answer affects your subsequent writing, revision, or opinion on the subject, it is considered correct and good academic practice to declare the use of generative AI as part of your scientific method.

Aalborg University's disciplinary rules regarding plagiarism apply at all exams!

During examinations or written tasks, where generative AI is a permitted aid, it is important to keep the rules regarding plagiarism in mind, and that you are not permitted to receive any unlawful assistance to solve the task. Even though the results of generative AI are often unique and original in their own right, the same rules apply to AI generated material as to other sources. This includes correct referencing and compliance with the rules of paraphrasing. The same applies to exam notes prepared with assistance from generative AI.

Matching disciplinary measures apply if generative AI is used to falsify or fabricate data, which is another example of violating the principles of good academic practice.

Due to the way generative AI models operate, even correct references to statements reflected through use of generative AI may be factually inaccurate. It is important that you pay attention to the occurrence of possible inaccuracies.

If you are unsure of whether or not you can recreate AI generated results (e.g. text), it is recommended as good practice to save as many pieces of information as possible regarding the time, the prompts, the output and version of the service/model. You can use screen-dumps to document this information.

Whenever you use generative AI to create images, videos, and audio, it is important to pay attention to current license terms and conditions, and to include correct information about this.

For example, it may concern material that is available under a Creative Commons license, which requires statement of use. Furthermore, you must comply with the service's terms of use that may contain further adjustments. Images must have correct references too.

Notice that the terms of the service includes the right to use your sources as input. This involves the risk of copyright infringement if the material you use is copyright protected or made available through special terms of use.

Students are NOT obliged to use generative AI as part of their education. Your teacher may recommend that you use it, but cannot demand that you use AI services other than those that have been GDPR approved. Currently those services are Microsoft Copilot and possibly some local AI models and services.

Whenever you use a service, it will handle your personal data. This includes the data you provide when you create your user profile, as well as more hidden data such as your location, the last homepage you visited, and which unit and browser you use etc. The provider uses the information to provide the service to you, and to improve and optimize the service.

Many providers of both free and paid-for services use your data for commercial purposes. This may involve profiling of you for tailored marketing purposes.

When you use generative AI, you should be aware that the provider can transfer your personal data to countries outside the EU e.g. the United States where the level of personal data protection may not be the same as in the EU due to GDPR.

Several generative AI services reserve the right to use your data to improve their services - including as training data. Therefore, you are permitted only to share information that is already made public, or that is free to be made publicly available.

You are NOT permitted to share other peoples' personal data e.g. name, gender, age, political beliefs, e-mails etc. with generative AI services. You are responsible for the data and must comply with GDPR. Sharing confidential information is prohibited. Information is considered confidential if it is based on e.g. collaboration with a company.

You can profit from reading more about generative AI and security, including a description of different services' data classification.