How can generative AI tools be used while studying?
Generative AI can support the study process; however, its use is subject to limitations specified by the University of Warsaw's URK Resolution and the guidelines of course instructors. After ensuring what scope of usage is permitted, we encourage you to explore the possibilities of these tools. Below, we present several examples of how generative AI can be used to create a personalized learning environment.
Generative AI as a support for learning:
The examples below come from the article by E. Mollick and L. Mollick (2023). We encourage you to read the full article. The authors explain in an accessible way how to formulate prompts appropriately, what to look for, and how to manage the risks associated with using these tools.
- AI as a mentor:
Generative AI can help create personalized, on-demand feedback/assessment, which in the case of interactions with a lecturer may be limited by working hours or group size. You can design the prompt so that the model “takes on the role” of the lecturer and helps you critically evaluate your work. It is a good idea to include in the prompt the purpose of the task for which you need feedback, and to specify the AI’s role, your expectations, and step-by-step instructions. Personalizing the prompt—taking into account the level of study, the detail of the suggestions you want, the level of work progress, etc.—can significantly improve the quality of the support you receive. Such support can take the form of a conversation with the AI, in which you can specify your needs. For an example prompt, see article by E. and L. Mollick (2023, p. 6-7).
- AI as a teacher:
You can also use generative AI to learn completely new topics, such as learning new research methods or familiarizing yourself with theories. In this case, the risk of getting incorrect information is greater, because the lack of experience in a given area can make it difficult to spot potential errors or “hallucinations” of the AI. Therefore, it is worth treating the AI as an additional source of information, complementing reliable materials. Remember that the AI’s answers may differ each time it is used. To get the best results, as with feedback, it is important to ensure that the prompt is detailed – specify the AI’s role, purpose, and provide clear instructions. It is worth asking for clear explanations, communicating what you have understood, and asking for examples, connections, or re-explaining in other words. See the example prompts in the article by E. and L. Mollick (2023, p. 12-14).
- AI as a simulator:
Generative AI can also support you in testing the knowledge you have already acquired by creating tests, scenarios, and cases to solve specific problems. Using generative AI in this way allows you to check whether you can apply your knowledge in both familiar and completely new contexts. When creating a problem simulator to solve, it is worth defining the context, the area of knowledge you want to practice, the roles that both you and the AI play, the characteristics of the scenario or test, and the way of interaction (see Mollick, Mollick 2023, p. 39).
- AI as a tool:
You can also use generative AI as a regular tool for things like coding, proofreading, translating text, summarizing and drawing key conclusions from text, responding to emails, etc., just remember to:
check your content,
track your usage and
cite generative AI tools appropriately.
Getting started with generative AI can be a challenge, as the results don’t always meet our expectations right away. However, it’s worth taking the time to get to know the technology – don’t be discouraged by the initial difficulties, but treat them as part of the learning process. The key to getting better results is to skillfully formulate prompts.
When working with generative AI, ethical issues such as privacy, copyright, and the risk of disinformation must also be taken into account. AI-generated answers may contain errors, even if they are presented in a trustworthy way. Therefore, the user is responsible for the final result of their work – when in doubt, it is worth reaching for information from reliable sources.
Generative AI as a support in conducting research: → see here
It is worth knowing that:
-
by logging in with your university email to your Microsoft 365 account, your daily limit in Microsoft Copilot is increased to 30 responses → link
-
GitHub Copilot offers a free program for students → link
-
depending on usage, access to OpenAI models (GPT-4, GPT-4o, DALLE, etc.) via API may be cheaper than subscribing to a Premium account → link
References:
Mollick, E. R., & Mollick, L. (2023a). Assigning AI: Seven Approaches for Students, with Prompts. SSRN Electronic Journal. DOI