4.1. Practical for the learner

AI applications support learning in various ways: they explain complex theories, provide ideas and feedback for research work, simplify statistical analysis and programming, and assist in translating, editing, and formatting text. However, the accuracy and quality of the final result remain the responsibility of the AI user.

AI applications can be used to solve various content-related problems. Here are some ideas on how AI applications could support your learning.

• Ask the AI application for examples related to the topic you are studying, the problem statement of your independent task (including thesis), or the research process. Let the AI application explain a complex theory in more straightforward terms or help you overcome a writing block (e.g., “Where should I start writing my thesis if I have a blank page fear?”).

• Use the AI application as a discussion partner to help you reflect on ideas – this is especially useful if you feel your questions might seem silly to your supervisor (although there are no silly questions). For example, ask for feedback on your research goal or research questions.

• Ask the text-based AI model for advice on synonyms you could use to find scientific sources. When inquiring about sources, be cautious, as this might sometimes lead to AI-generated (hallucinated) references. Especially with free AI applications, it is helpful to add clarification like “Recommend sources that exist.”

Never lose vigilance when working with AI. Always double-check the output received from AI to ensure the accuracy of the content.

• To find useful sources, you can always use the scientific information databases accessible through the University of Tartu Library. For more details on database usage, visit the Tartu University Library’s website and refer to the Information Search Principles (in Estonian).

• AI applications also assist in more technical tasks. Try asking for guidance on statistical methods and data visualisation techniques or clarifying and simplifying results interpretation. If necessary, AI applications can also help identify programming-related issues.

Text-based AI models can be helpful at various stages of working with text.

• Better AI applications can help you get an overview of an extensive source. However, never copy the information produced by the AI application directly into your work.

• AI applications can be used to translate text (parts). Various AI applications (e.g., Microsoft Copilot, ChatGPT, DeepL) can translate text from English to Estonian quite well, but stay alert, as automatic translations sometimes result in awkward sentence structures or inappropriate word choices. Remember that other applications, such as the University of Tartu’s open-source NeuroTranslation, developed by the Language Technology Research Group, can also be used for automatic translation. It is vital to be familiar with local applications and their advantages; for example, NeuroTranslation’s unique feature is the ability to translate into 23 Finno-Ugric languages.

• You can also try having the AI application edit your text, such as identifying word repetitions or shortening sentences. Always check the AI-edited text in the end: you are the author and responsible for the quality of the presented work.

• AI applications can also help learn academic expression. Let the text-based AI model phrase your response in different roles, specifying in the prompt, for example, “You are a scholar in the field of…” or “You are the supervisor of a master’s thesis.”

• AI applications can assist in formatting various parts of the text. For instance, let the text-based AI model format references correctly. You can also use AI to clarify the structure of your work, including correcting the distribution of structural sections.

Read more:

• Using AI in Writing (in Estonian) https://www.teadustekst.ut.ee/chatgpt-ja-teised-tekstirobotid/
• ChatGPT Recommendations for Learners https://openai.com/chatgpt/use-cases/student-writing-guide/

Task

Student Riina was assigned to provide feedback on a peer’s (Mari’s) work. Mari’s work included the following list of sources. Riina found it suspicious because she had worked through the list of sources recommended by the instructor for this task, and the sources presented in Mari’s work did not match that list. Riina wanted to give feedback on Mari’s sources, so she began by checking the accuracy of the sources Mari had provided. Did Riina find these sources?

Later, when Riina spoke with Mari, Mari said she had used AI to create her work and provided the following citation: “I am writing a paper on children’s language development, multilingualism in Estonia, children’s language acquisition: sociolinguistic and psycholinguistic perspectives, children’s language acquisition and development, language learning methods and learner development. Generate a summary of this work with references and a list of sources used. Use at least six sources.”

Try asking at least four different AI applications for relevant sources on your selected topic using a similar prompt. What sources do you get? Check if they are accurate and correctly formatted. What can you conclude from this?

Review the prompt creation recommendations provided in this material and analyse why the prompt in this task is ineffective.

Self-assessment