ChatGPT and other Forms of Artificial Generative Intelligence

ChatGPT and other forms of generative intelligence offer many new possibilities. However, be aware that you must not use it in connection with exams unless specifically allowed in the study programme or module description.

You are responsible for investigating whether the use of generative artificial intelligence is permitted in connection with your exams. If in doubt, you can always check in your study programme or module description.

When Using ChatGPT and Similar Tools in Exams where it is allowed

If it is permitted to use generative artificial intelligence for an exam, there are several things you need to be aware of.

Remember to cite your sources
If you use information generated by ChatGPT or similar tools, always make sure to provide proper citations. If you fail to reference the service you have used, it constitutes plagiarism, and Absalon will take plagiarism action against you.

Use your critical thinking
The answers you receive through generative artificial intelligence are not necessarily always true or comprehensive. It is important that you do not use the answers uncritically and that you use other tools to verify the validity of the answers. Be mindful that:

• You do not know the source of the answer

• The service can "hallucinate"

• The service may be subject to ethical codes that omit or suppress information.

You don't know the sources of the answer
When you receive answers via ChatGPT or similar tools, you are not aware of the extent of the material the service uses as its source. Often, the answers are based on text from news media, openly accessible research, social media, discussion forums, etc.

The service can "hallucinate"
Services like ChatGPT can hallucinate. This means they can invent sources or facts that are not accurate. This is due to the way these models operate, and to put it simply, the algorithm will try to guess an answer that sounds convincing, even if it lacks enough material to actually answer the question correctly.

Ethical codes
Many services that utilize generative artificial intelligence are subject to ethical codes that shape the answers to some extent. Therefore, there might be viewpoints in the service's source material that are suppressed or omitted in the service's responses.

If you use ChatGPT and similar tools in exams where it is not allowed

If you use ChatGPT or any other form of generative artificial intelligence in connection with exams at Absalon, it will be considered cheating.

If Absalon suspects that you have used ChatGPT or similar tools in an exam where it is not allowed, you will be summoned to a meeting with your Head of Study and an instructor. You are welcome to bring an advocate to this meeting.

During the meeting, you will need to explain how you prepared for the exam, how you worked on the assignment, and what considerations you made in connection with your exam response. The meeting will also include questions about the exam's academic content with the purpose of assessing whether you actually possess the knowledge and skills that match the exam response.

The meeting is not an exam, but it may have an exam-like character, as it aims to demonstrate your academic level and your knowledge of the current exam subject. This knowledge is an important element in the assessment of whether you cheated on the exam.

If Absalon determines that cheating has occurred, you have violated Absalon's Guidelines for good behaviour and therefore risk:

• receiving a warning

• being suspended for a period

• being permanently expelled.

What is generative artificial intelligence?

Generative artificial intelligence is a branch of artificial intelligence that deals with creating new content based on existing data. Using input from a human or a system, entirely new content can be generated, such as texts, images, and speech. The most well-known service is ChatGPT.