Generative Artificial Intelligence and Library Instruction: Ethics/Copyright

GenAI applications can sometimes be used as a tool for assignments (visual, writing and programming assignments). We expect students to build on this content and add argumentation, critical analysis, correct citation of sources, creative input, personal views or reflection, innovative character, connection with social context, etc. Make sure that you are able to discuss or have a debate about your assignment.

When people hear a sentence, they attach a meaning to it, connecting language to reality. A generative language model only has access to the format of sentences, but cannot connect to reality in any way. This means you have no guarantee that the system's texts are actually correct. The system's only concern is to generate texts that seem as plausible as possible; truth or reality don't matter to the model.

Use GenAI if it’s allowed, but don't blindly trust technology. The more responsibility you give the system, the more you will need to verify, check and explain the output. Sometimes, AI output seems very convincing but it’s perfectly possible that the answer is incorrect or even made up ('hallucinating AI'). In most cases, there’s no way to find out how the algorithms came to a particular result and there is no transparency about the sources the system used ('black box'). Carefully check the content you want to use and look for existing sources to refer to. After all, you are always responsible for the content of the assignment you submitted.

ChatGPT is known for generating fictitious references. GenAI's output sometimes lacks transparency about the sources the system used, increasing the likelihood of plagiarism. There is also a real risk of copyright infringement. GenAI builds on other people's work, and that always requires proper acknowledgement of the sources used.

ChatGPT is known for generating fictitious references. GenAI's output sometimes lacks transparency about the sources the system used, increasing the likelihood of plagiarism. There is also a real risk of copyright infringement. GenAI builds on other people's work, and that always requires proper acknowledgement of the sources used.

AI applications are trained with certain datasets, which are not always very representative. There is no transparency about the filters they apply, which raises a lot of ethical questions. If you blindly copy the output, you risk spreading incorrect information, certain stereotypes or prejudices.

There is little transparency about what developers do with the information fed to the system. Do not give personal data or confidential information to these GenAI applications. Check whether you have the necessary permission to enter copyrighted content. If you are in doubt about whether information is confidential or not, you can check with the provider of the information.

Know that the servers for AI applications have a very high energy consumption, so use them only when they can add value.

You can only show that you have learned something if you are transparent about using AI output and add your own contribution to the assignment. Keep in mind that if you are not transparent, this can be considered an irregularity and you risk a penalty. If there are no clear guidelines on this topic in the assignment description, ask your teaching staff about it.

Talk with the faculty and instructors of the departments you serve to understand and adhere to the policies they have regarding the use of GenAI.