GenAI roles in your course

Rationale

Writing essays that meet academic standards equips students with the skills necessary for clear communication, critical analysis, and argumentation. These competences are essential for succeeding in academic environments. GenAI can assist by offering writing guides, providing feedback on structure and grammar, and ensuring adherence to academic conventions, thereby enhancing students’ ability to craft well-organised and authoritative essays.

Learning objective

Students can write essays that meet academic standards.

GenAI role

The LLM acts as a specialised academic writing tutor, providing personalised feedback on essay structure, argumentation, style, and grammar.

Important note: While the GenAI persona mimics an academic tutor, it is still a machine outputting statistically likely language. Students must approach the feedback critically and verify its reliability. Class discussions are vital to support such critical thinking.

Lecturer role

During the class discussions, lecturers can provide their own tips and guidance, and help students reflect on the GenAI-generated feedback they received. Any feedback shared by the students should be considered as starting points for a more generalised, lecturer-led discussion.

Learning activities

Students write weekly essays and use UvA AI Chat to receive feedback from a GenAI persona. They submit the prompts used alongside their drafts and reflect on the quality and limitations of the GenAI’s feedback. During seminars, both writing and prompting are discussed with the whole class. Optionally, students share their chat logs.

Instructions for implementation

When using this activity, please consider the following:

• Before the course starts, make sure all students are in a single UvA AI Chat group.
• Share good example prompts for feedback requests with this class group.
• Decide on whether you want students to share prompts or full conversations with the class.
• Create Canvas assignments where students can submit their weekly essays, as well as prompts and/or chat logs.

Assessment

• Weekly drafts can be assessed formatively to ensure alignment between student and lecturer.
• Contribution to class discussions can provide a measure of cognitive engagement with the learning process.
• A final essay can be the summative measure of the student’s ability to write an academic essay.
• The quality of interaction with GenAI feedback, on the basis of chat logs, can provide a measure of AI literacy.
• To assess learning skills, students can be asked to reflect on their use of GenAI prompts, including how and why they acted on feedback from the persona.

Variations

• Have an in-class session in which a student receives both peer feedback and GenAI feedback. Have them look at the differences: Which feedback was more useful?
• Rather than individual essays, have students write collaborative essays. They use a collaborative writing tool and track changes to show how they processed GenAI feedback on their writing. They then evaluate the collaborative writing process and GenAI feedback during group discussions in class.

Tips

• Students:
  • Understand the why behind GenAI’s suggestions.
  • Agree before implementing.
• Lecturers:
  • Use interesting GenAI output for a class discussion.
  • Provide clear guidelines for GenAI use in essays.
  • Discuss with students how GenAI personas might lead to overconfidence in flawed advice.
  • Encourage students to validate feedback before implementation.
• Reflection question: Which prompts or strategies did I find most effective for getting useful feedback?

Example at the Science Faculty

At the bachelor programmes Psychobiology and Biomedische Wetenschappen, students were allowed to use GenAI as a writing coach when writing their review paper and mini thesis. Visit this page on the TLC website for more information about these two pilots.

Rationale

Formulating a research question is a key academic skill. Quality research depends on questions that are clear, anchored, not too narrow or broad, relevant, researchable and analytical. Learning how to formulate a good question is a matter of calibration. Students can use GenAI to get direct feedback to perform this calibration, as long as they focus on receiving feedback rather than generation by GenAI.

Learning objective

Students can formulate a research question and revise it in response to structured feedback.

GenAI role

The LLM acts to address specific questions about research questions offered by the students. It provides constructive advice to help students improve their work.

Lecturer role

Before the brainstorming session, the lecturer should make sure to convey the key aspects of a good research question and the value of iteration. The lecturer should also show some examples of good research questions and of questions that need more refinement.

Learning activities

Before using GenAI, have students develop a preliminary research question (in class). Once they have a first version, the students can start refining their research question with the help of GenAI. Students are tasked to document the full refinement process.

Instructions for implementation

• Make sure all students are placed in a single UvA AI Chat group.
• Develop a persona that guides students through the refinement process of a research question and add this persona to the UvA AI Chat group. If you develop your own, test it a few times with research questions of differing quality.
• Clearly instruct students to first develop their question as well as they can, so they receive GenAI feedback on their best effort and can learn optimally. Similarly, convey that GenAI output should not be considered authoritative and that they should only incorporate feedback that seems sensible to them.

Assessment

Students submit their initial research question and each subsequent revision, along with a reflection explaining the changes made and the rationale behind them. The assessment rubric considers the quality of the final research question and the thoughtfulness of the iterative refinement process.

Variations

• Conduct a group brainstorming session where students share their preliminary research questions and receive feedback from their peers. Encourage them to consider different perspectives and potential areas of exploration for their chosen topic.
• Have students conduct a mini literature review on their chosen topic before refining their research question. This will help them understand the existing knowledge and identify gaps that they can address with their research.
• Assign students to work in pairs to refine their research questions, allowing them to collaborate and provide feedback to each other as they develop their ideas.

Tips

• Students:
  • Be specific in your requests for feedback from the LLM. For example, instead of asking “Is this a good question?”, ask “Is this question too broad? How can I make it more focused?”
• Lecturers:
  • Model the process of refining a research question by showing examples of weak and strong versions.
  • Provide clear criteria (specific, researchable, relevant, analytical) that students can use when judging the quality of their own and others’ questions.
  • Encourage students to compare GenAI feedback with feedback from peers and course materials.
  • Highlight the importance of iteration: Reassure students that a good research question usually emerges after several revisions.
  • Use class time to discuss common pitfalls (too broad, too narrow, unfocused) and show how the LLM’s suggestions can help, but also mislead, students.
• Reflection question: How did the LLM feedback change your understanding of your research question?

Example at the Science Faculty

At the bachelor programme Future Planet Studies, students could use a persona that gave targeted feedback on their research question. Visit this page on the TLC website for more information about this pilot and visit this page for the system prompt that was used for this persona.

Rationale

Transforming informal lab notes on experiments into formal, publishable-quality method sections enables students to document their procedures and enhance the credibility of their research. This skill is necessary for contributing to scientific literature and ensuring reproducibility. GenAI can support this by offering templates, providing editing assistance, and suggesting improvements, helping students convert their detailed observations into clear, professional method sections ready for publication.

Learning objective

Students can produce method sections based on experimental lab notes that meet disciplinary standards.

GenAI role

The LLM acts as a writing assistant, converting student lab notes into a structured and formalised method section draft.

Lecturer role

Before the practical, students should be instructed about the level of detail that is required for a good lab journal and how to take notes effectively. During the practical, lecturers can guide the process if students have never maintained a lab journal before.

Learning activities

During practicals, students make notes in their lab journal as usual. After the experiment, they submit their notes into UvA AI Chat and prompt it to generate a method section, which they then revise. They submit both the generated and revised version, to provide transparency about their process.

Instructions for implementation

• Make sure the student group has access to UvA AI Chat before the course starts.
• Optionally, develop a persona that processes the lab journal notes using a fixed template for a method section.
• Develop clear grading rubrics that detail what the hallmarks of a good method section are.
• Experiment with the performance of GenAI before implementing this activity into your course.

Assessment

Evaluate the student’s final revised method section for clarity, accuracy, completeness, and adherence to scientific writing conventions.

Variations

• Some GenAI applications are multimodal, meaning that the user can also submit pictures. You can experiment with providing such pictures rather than written notes to generate a method section.
• You can task students to provide good context to the GenAI system before running the experiment transferring the notes. The added benefit of this would be that they engage with the rationale behind the experiments and prepare for their tasks.

Tips

• Students:
  • Carefully review and edit the LLM-generated draft.
• Lecturers:
  • Provide examples of high-quality method sections.
  • Emphasise the LLM is a tool, not a replacement for understanding.
• Reflection question: How did using the GenAI tool change my approach to writing a formal method section?

Example at the Science Faculty

At the bachelor programme Scheikunde, students could use UvA AI Chat to prepare their laboratory experiment and they could use a persona to receive feedback on all parts of their practicum report. Visit this page on the TLC website for more information about this pilot and visit this page for the system prompt that was used for this persona.

Rationale

Using GenAI to improve understanding of programming concepts and problem-solving helps students enhance their learning efficiency and tackle complex coding challenges more effectively. This ability is essential for advancing their programming skills and developing innovative solutions. GenAI can support this by offering interactive tutorials, providing instant feedback, and suggesting optimised code solutions, thereby enabling students to grasp concepts more thoroughly and solve problems with greater proficiency. GenAI is able to do this with many programming languages.

Learning objective

Students can use GenAI tools to generate, assess, and improve code solutions, and explain the reasoning behind algorithmic choices.

GenAI role

• Code generation and explanation: LLMs can generate code snippets for specific functionalities and explain existing code, clarifying the purpose of individual lines or functions.
• Debugging assistance: Students can describe errors they encounter to the LLM, which can suggest potential causes and solutions. LLMs can also analyse code for potential bugs or inefficiencies.
• Concept clarification: Students can ask the LLM to explain specific programming concepts relevant to the exercises or project (e.g., “Explain the concept of a for loop in Python”).

Lecturer role

• Exercises: Running the exercises in class will allow the lecturer to help students directly or to help them with assessing LLM output. This way, common difficulties can be subject of plenary discussion and students can be encouraged to not outsource work to UvA AI Chat, but rather to use it to further their understanding.
• Mini-Project: The lecturer can focus on good coding practices, documentation and the ability of students to justify their decisions.

Learning activities

• Exercises: Students work through a series of programming exercises focused on domain-specific concepts. They can use UvA AI Chat to give example code with explanation, help with troubleshooting code, and explain programming-related concepts.
• Mini-Project: Students undertake a small-scale scientific programming project applying learned concepts to a real-world problem. They can use UvA AI Chat for debugging assistance.

Instructions for implementation

• Make sure the students are added to a UvA AI Chat group.
• Develop a persona that can coach students through programming questions and add this persona to the UvA AI Chat group. You can provide the persona with specifics about the course learning objectives and student backgrounds.
• Make sure all staff overseeing the programming practicals have experimented with using UvA AI Chat.

Assessment

• Exercises: Code reviews focusing on readability and adherence to best practices. Short quizzes or reflections on programming concepts.
• Mini-Project: Project report outlining the problem, approach, code implementation, results, and discussion. Code review and presentation of the project.

Variations

• Rather than a Mini-Project, offer a timed case study for assessment. Students analyse and solve programming challenges in specific scientific fields, such as predicting weather patterns or analysing genetic data. They can use UvA AI Chat to discuss their approaches and receive feedback on their solutions.
• Peer teaching: Students take turns teaching each other programming concepts and applying them to scientific problems, such as explaining how to implement machine learning algorithms for analysing medical data. They can use UvA AI Chat to ask questions and clarify any misunderstandings during the teaching process.

Tips

• For students:
  • Focus on understanding the underlying concepts and using the LLM as a tool for learning and problem-solving.
  • Document how you used the LLM in your project and reflect on its effectiveness.
• For lecturers:
  • Offer opportunities for students to share their experiences and best practices for using LLMs in scientific programming.
• Reflection question: How did the LLM enhance my understanding of programming concepts and problem-solving strategies?

Example at the Science Faculty

At the master programme Physics & Astronomy, students could use UvA AI Chat to help them with programming homework exercises. Visit this page on the TLC website for more information about this pilot.

Rationale

Assessing and selecting statistical techniques is important practice for students. It bring them to think critically about the research design and aims, as well as the strengths and limitations of different statistical methods. This promotes a deeper understanding of the research process and the ability to make informed decisions. In addition, by mastering this skill, students are better prepared for future research endeavours and can contribute meaningfully to their respective fields.

Learning objective

Students can assess and select statistical techniques that align with the design and aims of a research project.

GenAI role

The LLM acts as a brainstorming partner, offering initial suggestions for statistical methods based on the project description.

Lecturer role

Prior to the selection of appropriate methods, the students need to gain an understanding of statistics and the various statistical methods on offer. The lecturer provides such knowledge and can guide students through the decision-making steps during the writing process.

Learning activities

During a series of seminars, students develop a research proposal on the basis of a research question, detailing experimental design and relevant hypotheses. They can then use UvA AI Chat to provide suggestions for statistical analyses, which they evaluate using their own understanding of statistics. Their task is to select a proper analysis method and justify their choice.

Instructions for implementation

• Ensure students have access to UvA AI Chat before commencing the course.
• Prepare some example conversations which you can share with students as an example of effective prompting for the purposes of this learning activity.
• Clarify that students should be able to justify their decisions (why this and not that?) and develop an appropriate grading rubric.

Assessment

Students submit a written justification of their chosen statistical method that:

• References GenAI suggestions.
• Demonstrates understanding of appropriateness.
• Critically analyses alternative methods.

The assessment rubric should assess clarity, depth, and quality of justification.

Variations

• Case study analysis: Students are given a real-world research case study and are tasked with developing a research proposal and selecting appropriate statistical analysis methods using UvA AI Chat. They present their choices and justifications to the class.
• Interactive workshop: Students participate in an interactive workshop where they use UvA AI Chat to explore different statistical analysis options for a given research question. They then engage in group discussions to select and justify the most suitable analysis method.
• Peer review activity: Students exchange their research proposals and GenAI-generated statistical analysis suggestions with a peer. They provide feedback and suggestions for improvement, including justifications for their chosen statistical analysis methods.

Tips

• Students:
  • Be detailed with project description; research GenAI suggestions thoroughly.
• Lecturers:
  • Provide guidelines for effective GenAI prompts. For more information, visit this page for our prompt engineering tips.
  • Offer additional resources (textbooks, articles, tutorials).
• Reflection question: How well did the LLM’s suggestions align with my own initial ideas about appropriate statistical methods for my research project?

Rationale

Formulating and refining plausible hypotheses about a case study helps students develop critical thinking and analytical skills by justifying their reasoning with evidence. This process is essential for constructing sound arguments and conducting rigorous research. GenAI can assist by providing access to relevant data, expert opinions, and analytical tools, enabling students to build well-supported hypotheses and refine them effectively.

Learning objective

Students can formulate and refine hypotheses about a case study, justifying their reasoning with evidence and expert insights.

GenAI role

The LLM can be used to explore scenarios suggested by the students, offering potential alternative explanations or counterarguments based on the provided evidence. In essence, it can be used to help the students think critically about their own thought processes.

Lecturer role

To prepare students for this complex task, it is advised to run through one or more guided examples, modelling the decision-making process from case description to substantiated, testable hypothesis.

Learning activities

Students are presented with a case study (e.g., a crime scene, patient description, business case). They select and organise data they deem to be evidence, brainstorm scenarios and then use the LLM to challenge their assumptions and explore alternative scenarios. On the basis of the most plausible scenario, they create a testable hypothesis.

Instructions for implementation

• Ensure that students have access to UvA AI Chat before the course commences.
• Optionally, create distinct personas for the different phases: To challenge the scenarios created by the students, to explore alternative ones and possibly to refine a hypothesis based on the selected scenario.

Assessment

Assessment can be done through written reports, oral presentations, or class debates. A rubric focusing on critical thinking skills, evidence evaluation, and reasoned argumentation should be used.

Variations

• Role-play simulation: Students are assigned different roles within the case study (e.g., detective, doctor, business manager) and are tasked with gathering evidence and brainstorming scenarios from their unique perspectives. They then use UvA AI Chat to challenge their assumptions and explore alternative scenarios before presenting their findings to the class.
• Multimedia analysis: Students are provided with multimedia resources related to the case study (e.g., video footage, audio recordings, documents) and are asked to extract and organise evidence from these sources. They then use UvA AI Chat to analyse the evidence and generate alternative scenarios, ultimately creating a testable hypothesis based on their findings.
• Visual mapping activity: Students use visual mapping tools to organise and connect different pieces of evidence from the case study, creating a visual representation of the various scenarios they brainstorm. They then use UvA AI Chat to challenge their assumptions and explore alternative connections between the evidence, ultimately leading to the development of a testable hypothesis.

Tips

• Students:
  • Carefully select the evidence you wish to use before engaging with UvA AI Chat. Clearly document this and your interactions with UvA AI Chat to demonstrate your own reasoning process.
• Lecturers:
  • Consider having students evaluate the LLM’s responses for accuracy and bias.
• Reflection question: How can I evaluate the biases and assumptions in the LLM when it proposes scenarios?

Rationale

Evaluating scientific arguments and identifying weaknesses and underlying assumptions empowers students to deepen their understanding and enhance their analytical skills. This capability is essential for evaluating the validity of research and constructing (robust) scientific work. GenAI can assist by highlighting logical flaws, suggesting areas for further investigation, and providing tools for in-depth analysis, helping students accurately assess and strengthen scientific arguments.

Learning objective

Students can analyse scientific arguments and identify weaknesses and underlying assumptions.

GenAI role

The LLM provides targeted feedback on student-generated argument maps in the form of critical questions, prompting students to reflect on their analysis and identify potential flaws or gaps.

Lecturer role

During an in-class activity, demonstrate and practice with the generation of simple argument maps. Also practice with submitting an argument map to an LLM.

Learning activities

Students read scientific articles, create argument maps to visually represent the reasoning, and then use an LLM persona to critique the argument. This does not have to be done in-class, but can be done during self-study, while time in class is spent on discussing the results.

Instructions for implementation

• Make sure all students are placed in a single UvA AI Chat group.
• Develop a persona that critiques or critically analyses arguments maps and add this persona to the UvA AI Chat group.

Assessment

Have students present a critical assessment of a scientific article, for which they used a self-created argument map.

Variations

• Instead of creating argument maps, students can listen to scientific podcasts and then engage in discussions using an LLM persona to critique the arguments presented in the podcast.
• Students can analyse real-world case studies related to scientific research, create argument maps to visually represent the reasoning, and then use an LLM persona to critique the arguments and propose alternative solutions.

Tips

• Students:
  • Clearly articulate the argument in your map before submitting it to the LLM.
• Lecturers:
  • Provide clear guidelines for argument mapping and examples of effective critiques.
• Reflection question: How effectively did the LLM-generated feedback help me refine my argument map and deepen my critical analysis of the scientific article?

Rationale

It’s important for students to learn to analyse a debate topic from multiple perspectives and formulate nuanced arguments and counterarguments. This fosters critical thinking, improves their ability to understand complex issues, and enhances their persuasive communication skills. GenAI can assist by presenting diverse viewpoints, providing frameworks for structuring arguments, and suggesting relevant evidence, helping students develop well-rounded and compelling debates.

Learning objective

Students can analyse a debate topic from multiple perspectives, formulating nuanced arguments and counterarguments.

GenAI role

The LLM acts as a sparring partner, providing arguments and counterarguments from various perspectives based on assigned personas. It helps students explore the breadth and depth of the debate topic beyond their initial understanding.

Lecturer role

The lecturer helps students prepare for the debate. Depending on their level, this can include classes on logical argumentation and rhetoric, or classes that discuss the content of the resources which will be used during the debates.

Learning activities

Besides general debate preparation, organise a class in which the students develop their arguments for the debate. To ensure they remain cognitively engaged, make use of distinct stages:

• Researching and identifying arguments without LLM use.
• Considering the viewpoints of relevant stakeholders without LLM use.
• Revising their argumentation.
• Receive feedback on their argumentation (LLM allowed).
• Practice simulated discussion with distinct personas (LLM-based).
• Revising their argumentation.

Instructions for implementation

• Make sure all students are placed in a single UvA AI Chat group.
• Develop different personas, corresponding to the different viewpoints, and add these personas to the group.
• Practice some debates with each of the personas, to see if they function as intended.
• Add some example prompts for the preparatory stages, in which students can use the personas to get feedback on their argumentation or look for new angles on their debate topic.

Assessment

• Debate performance: Evaluate students’ ability to articulate nuanced arguments, anticipate and refute opposing viewpoints, and effectively incorporate diverse perspectives into their debate strategy.
• Argument map: Students create a visual map illustrating the different perspectives generated by the LLM, highlighting key arguments and counterarguments for each persona.

Variations

• Instead of a traditional debate, have students participate in a role-play simulation where they take on the personas of different stakeholders after practicing interactions with the LLM.

Tips

• Students:
  • Do your research and map your arguments before engaging with the LLM personas.
• Lecturers:
  • Provide a list of potential personas relevant to the debate topic to help students get started.
  • Encourage students to fact-check information.
• Reflection question: How plausible was the ‘simulation’ offered by the LLM?

Example at the Science Faculty

At a bachelor course at Amsterdam University College, students could practice their debating skills by using four different personas, each with their own ideological beliefs. Visit this page on the TLC website for more information about this pilot and visit this page for the system prompts that were used for these personas.

Rationale

It is important for students to learn to analyse and articulate the perspectives of diverse stakeholders in real-world project scenarios because it enhances their understanding of complex, multifaceted issues. Acquiring this skill ensures they can navigate and address varying interests effectively. GenAI can support this learning by providing tools for stakeholder analysis, facilitating the synthesis of diverse viewpoints, and offering examples of effective communication, thereby enriching students’ analytical and articulation abilities.

Learning objective

Students can analyse and articulate the perspectives of diverse stakeholders involved in a real-world project scenario.

GenAI role

The LLM acts as a ‘perspective simulator’. Students provide the LLM with background information about the project and a specific stakeholder group. They can then prompt the LLM to:

• Generate possible questions the stakeholder might ask.
• Predict the stakeholder’s potential reactions to different project proposals.
• Simulate a conversation with the stakeholder, allowing students to practice their interview skills and refine their approach.
• Summarise key concerns and priorities expressed by the simulated stakeholder.

Lecturer role

During projects, the lecturer usually takes on a coaching role. Within this activity, this coaching role is two-fold: To help the students empathise with the stakeholders (a role similar to that played by the LLM) and also to guide the students who are using an LLM through the process, by briefly joining in with them to check prompts, responses and interpretation.

Optionally, the lecturer could also teach the students about interview techniques before having them work with the LLM. This way, they are also prepared for interviews with real people.

Learning activities

In preparation for the class activity, ask students to write a summary of their project and identify key stakeholders. During class, let them develop personas or prompts so that LLMs will roleplay these stakeholders and ask students to pitch their project, engage in dialogue and record the responses.

In a plenary activity, have the different students groups share their summary, stakeholders and stakeholder responses with the group. Discuss the new insights and encourage peer feedback among students.

After this activity, task the students to reach out to actual stakeholders for an interview.

Instructions for implementation

• Make sure all students have access to UvA AI Chat before the course starts.
• Develop personas or prompts to simulate a few common stakeholders and practice the assignment: Do the simulated stakeholders respond realistically?

Assessment

• Interview performance: Evaluate students’ ability to conduct effective interviews with actual stakeholders, demonstrating active listening, thoughtful questioning, and respectful communication.
• Stakeholder feedback: If feasible, gather feedback from the actual stakeholders on the students’ understanding of their perspectives and the effectiveness of the project.

Variations

• Have students form panels representing different stakeholder groups (optionally after practicing with an LLM) and engage in a discussion about the project. This will allow for a more dynamic and interactive exchange of ideas and perspectives.
• Ask students to create visual maps of their project’s stakeholders, including their relationships and levels of influence. This will help them better understand the stakeholder landscape and identify potential areas of support or resistance.

Tips

• Students:
  • Clearly define the stakeholder group and their relationship to the project before prompting the LLM.
  • Don’t rely solely on the LLMs output. Use it as a starting point for your own research and critical thinking.
• Lecturers:
  • Encourage students to compare and contrast the perspectives generated by the LLM with information gathered from other sources.
  • Offer support and guidance on effective interviewing techniques.
• Reflection question: How did engaging with the LLM help me to better understand the perspectives of stakeholders, and where did I still need to adjust or refine my own assumptions?

Rationale

Students need the skills to develop and apply a Team Charter because it fosters clear communication, defines roles and responsibilities, and ensures accountability, all of which are essential for effective collaboration in group projects. GenAI can support this by facilitating structured discussions, providing templates, and tracking adherence to agreed-upon norms, thereby enhancing teamwork, responsibility management, and conflict resolution abilities crucial for success in academic and professional environments.

Learning objective

Students can create a Team Charter to take shared responsibility during a group project.

GenAI role

The LLM acts as a ‘virtual team coach’. It provides feedback to the group, highlighting both strengths and areas for improvement, and suggests strategies for enhancing team performance based on the Team Charter.

Lecturer role

The main role of teaching staff is to explain and validate the concept of a Team Charter and to monitor whether students succeed in drafting their own. Then, lecturers can instruct students to submit the charter to UvA AI Chat.

Learning activities

The main learning activity is drafting a Team Charter. Prior to this activity, each student creates a personal ‘user manual’, summarising what their team mates should know about them. During the class activity, students draft an agreement between their team members, listing goals, values, norms for behaviour during the project, strengths of the team and potential weaknesses, as well as the purpose for the team.

This charter can then be submitted to UvA AI Chat by one (or all) of the students, together with the following prompt:

“We are a group of students working on a group project. Attached you will find our Team Charter. Each week we will report our progress in this chat. Please provide us with constructive feedback on the basis of our progress and the Team Charter.”

During the project, the team can document their project progress and receive tailor-suited advice from the LLM.

Instructions for implementation

• Make sure all students have access to UvA AI Chat before the course starts.
• To structure the development of a Team Charter, offer a template to the students.
• Do a practice run with a mock Team Charter and simulated progress.
• Check in with the students on a weekly basis and ask them about their progress and feedback.
• Instruct students to be transparent about their progress and also mention frustrations/difficulties, in order to get better feedback. Emphasise that the chat logs are not accessible to you.
• Be sure to still organise progress meetings and do not delegate project group monitoring to the LLM.

Assessment

• Reflection: Students individually and collectively reflect on their team’s performance, identifying successes, challenges, and lessons learned regarding teamwork and the effectiveness of the Team Charter. This reflection can take the form of a written report, presentation, or class discussion.

Variations

• Instead of creating a Team Charter, students could conduct a personal strengths and weaknesses analysis and share it with their team members. They can then work together to identify how they can leverage each other’s strengths and support each other’s weaknesses during the project.
• Instead of submitting the Team Charter to UvA AI Chat, students could engage in a peer feedback exchange. Each week, they could provide constructive feedback to their team members based on their progress and the agreed upon Team Charter. This would promote a culture of continuous improvement within the team.
• Students could keep a project progress reflection journal where they document their individual and team progress, challenges faced, and lessons learned. They can then use this journal as a basis for discussions with their team members and UvA AI Chat to receive feedback and guidance throughout the project.

Tips

• Students:
  • Be specific and detailed in the Team Charter.
  • Regularly check the LLMs reports and discuss them as a team.
  • Consider sharing progress reports with the LLM.
• Lecturers:
  • Model the process of creating a Team Charter.
  • Ensure students understand the purpose and value of the LLMs feedback.
  • Facilitate discussions around the LLMs insights, helping students interpret the data and develop actionable strategies.
• Reflection question: Did the LLM’s feedback help us identify and address areas for improvement? How?

Rationale

Students developing (and sticking to) personalised learning plans that align with course objectives is important because it empowers them to take ownership of their learning, tailor their study strategies to their individual needs, and effectively achieve their academic objectives. GenAI can assist by offering tools for goal setting, tracking progress, and providing personalised recommendations, enhancing the customisation and efficiency of students’ learning experiences.

Learning objective

Students can develop personalised learning plans aligned with course goals.

GenAI role

The LLM acts as a personalised learning assistant:

• Explaining and comparing various learning strategies (e.g., spaced repetition, active recall, Feynman technique).
• Helping students structure their learning plan, including timelines, milestones, and specific activities.
• Offering prompts to encourage students to reflect on their learning process and identify areas for improvement.

Lecturer role

The activity does not necessitate in-class work, although it can be beneficial to discuss important learning strategies in class before asking students to engage with an LLM. It is also recommended the lecturer develops and tests useful prompts for the students.

Learning activities

It is possible for students to engage with the LLM to receive personalised feedback without an in-class activity, but for this it’s important to clearly instruct them how to do so, especially if you choose to assess their engagement.

You can opt for an in-class activity, to explain to them which information they should submit (e.g. learning objectives, deadlines for assignments, strengths and weaknesses), discuss learning strategies and practice a first attempt at getting useful feedback.

Instructions for implementation

• Make sure all students have access to UvA AI Chat and place them in a class group.
• Develop relevant prompt and personas, and share these with the group.
• Collect useful resources on learning strategies and share these on Canvas.

Assessment

Students can maintain a journal documenting their learning process and plan adjustments. The quality of their plans and adaptations can be considered a measure of their learning skills.

Variations

• Ask students to create a short video reflecting on their learning objectives, deadlines for assignments, and self-assessment of strengths and weaknesses. They can then submit the video for personalised feedback from the instructor.
• Facilitate a peer feedback exchange where students can partner up and exchange their insights about crafting a learning plan. They can then provide constructive feedback to each other and discuss strategies for improvement.

Tips

• Students:
  • Regularly review and revise your learning plan.
  • Actively reflect on your learning process and document your insights in your journal.
• Lecturers:
  • Model the process of creating and using a learning plan.
  • Emphasise the importance of self-reflection and metacognition.
• Reflection question: Are there any learning strategies suggested by the LLM that I did not trust or find useful? Why?

Rationale

Understanding how to critically evaluate research progress, identify strengths and areas for improvement, articulate ambitions, and reflect on ethical considerations and personal values are important skills when doing research. These skills ensure a thorough, reflective, and principled approach to research, which is essential for academic success. GenAI can support students by offering analytic tools to assess research data, suggesting improvements, facilitating the articulation of goals, and guiding considerations on ethical implications and personal values.

Learning objective

Students can assess their research progress, identifying strengths, weaknesses, and formulating goals for improvement and future direction.

GenAI role

The LLM acts as a personalised research mentor, posing questions to guide student reflection and personal growth. The LLM will encourage the student to draw on experiences in order to simulate future actions (prospective thinking).

Lecturer role

In the context of a research project, the supervisor will be a sounding board for progress, too. This role should not be diminished because of LLM use, but should rather complement it.

Learning activities

No in-class activities are necessary for this, although one might set up a class to explain the purpose of reflective research journaling and practice with using UvA AI Chat to do so.

Instructions for implementation

• Make sure all students have access to UvA AI Chat and place them in a class group.
• Develop and test a persona that coaches students during journalling and add it to the UvA AI Chat group.
• Set up clear instructions for the research journaling assignment, clarifying which topics should be addressed.
• Optionally, develop and test prompts you can share with the group.

Assessment

Students can write a reflection summary to identify what they learnt about themselves and research praxis, and how they plan to put their new insights to practice.

Variations

• The LLM and student switch roles, with the student posing reflective questions about the research project, after detailing the setup.
• The LLM encourages the student to share personal experiences through storytelling, and then helps the student analyse and draw insights from these stories to inform future actions.
• The LLM guides the student through a creative visualisation exercise, where the student imagines themselves in future scenarios and reflects on how they would handle different challenges based on past experiences.

Tips

• Students:
  • Approach the LLM as a thinking partner, be honest in your responses, and revisit past conversations to track your growth.
• Lecturers:
  • Emphasise the importance of authentic reflection and the private nature of the LLM chat.
• Reflection question: How did my interactions with the LLM shape my understanding of its capabilities?

Rationale

Being able to evaluate the relevance of key concepts to broader themes in the course helps students understand the interconnectedness of ideas, reinforce their learning, and deepen their comprehension of the subject matter. This skill enables them to apply their knowledge more effectively across different contexts. GenAI can assist by providing context, highlighting connections between concepts and themes, and offering insights and examples, fostering a more integrated and comprehensive learning experience.

Learning objective

Students can analyse the connections between core concepts and overarching course themes.

GenAI role

The LLM persona is directed towards course content and learning objectives. It prompts students to make connections between concepts, challenges their understanding, and provides feedback. It acts to elicit deeper thinking from the students.

Lecturer role

Since this is a self-study application, the lecturer role is limited. The lecturer can instruct students on how to do this and optionally practice in class. Besides this, the development of the persona (including the documents it references) should be offered by the lecturer. It is recommended to make personas for each week of progress, using a cumulative knowledge base.

Learning activities

This is a tool for self-study. Students can be instructed to make use of the LLM, but they should realise it’s an option, not a necessity. Mentioning the existence of UvA AI Chat and the weekly personas during the first class may be sufficient in-class attention.

Instructions for implementation

• Make sure all students have access to UvA AI Chat and create a class group.
• Develop and test the persona(s) and add them to the UvA AI Chat group.
• Do regular check-ins with the students on how useful the LLM feedback has been.

Assessment

Student understanding can be tested in any appropriate form (e.g. written exam).

Variations

• Instead of just prompting students to make connections between concepts, have them create a visual concept map that illustrates the relationships between different course concepts and learning objectives. This will challenge their understanding and provide a tangible representation of their learning.
• Only use the LLM in-class, so that you can monitor output and resolve any confusions. This can be part of a bigger seminar exercise to dive more deeply into a topic.

Tips

• Students:
  • Start with a narrow focus on connecting two specific concepts before expanding to larger course themes.
• Lecturers:
  • Encourage students to compare LLM feedback with their notes, textbooks, and peer discussions. Also remind them that the LLM’s suggestions should be verified against reliable course materials.
  • Use class time to let students share and critically discuss insights they gained from the LLM.
• Reflection question: Did the GenAI help you identify connections you hadn’t considered before, or did it reinforce your existing understanding?

Rationale

Developing plausible, substantiated future scenarios allows students to anticipate potential challenges and opportunities, think critically about the implications of current trends, and plan strategically. This skill enhances their ability to forecast and innovate. GenAI can support this by analysing data trends, providing simulation tools, and offering predictive insights, helping students create well-founded projections and scenarios.

Learning objective

Students can construct future scenarios informed by data, trends, and theoretical frameworks.

GenAI role

The LLM acts as an idea generator, brainstorming partner, research assistant (finding relevant information), and organisational tool (outlining, summarising, and structuring arguments).

Lecturer role

The lecturer facilitates the process by providing frameworks for scenario development and examples of well-structured scenarios. They guide students in critically evaluating the LLM output, ensuring that speculative ideas remain grounded in evidence and theory.

Learning activities

Students identify key trends and uncertainties in their field, then prompt the LLM to generate different scenario outlines. They compare, refine, and expand on these with additional research, eventually developing 2-3 plausible scenarios. In class, groups present and critique each other’s work.

Instructions for implementation

• Ensure students have access to UvA AI Chat.
• Provide templates for scenario development.
• Offer example prompts for how to generate balanced scenarios and remind students to document their prompting process.

Assessment

Assess the final product as a written report, oral presentation or visual representation, focusing on:

• Scenario plausibility
• Depth of analysis
• Coherence of arguments
• Effective use of evidence and supporting details

Variations

• Instead of written scenarios, students create podcasts or short videos presenting their future scenarios.
• Conduct a live “scenario workshop” where the LLM plays the role of a facilitator, challenging assumptions and prompting new directions.
• Combine with roleplay: Assign students different “future stakeholders” and let them debate the implications of the scenarios.

Tips

• Students:
  • Experiment with different prompts to explore diverse perspectives.
  • Don’t rely solely on the LLM.
  • Critically evaluate the generated content.
• Lecturers:
  • Offer prompt engineering workshops.
  • Facilitate in-class discussions about the scenarios.
• Reflection question: Did engagement with the LLM promote my own creativity and problem-solving?

Example at the Science Faculty

At the IIS honours course, students could use UvA AI Chat as a brainstorming partner for developing scenarios for the future educational landscape. Visit this page on the TLC website for more information about this pilot.

Rationale

Evaluating the effectiveness of a prototype solution against self-defined benchmarks helps students develop critical thinking and assessment skills, ensuring that their solutions meet specific criteria and achieve desired outcomes. This ability is essential for refining projects and enhancing their practical applications. GenAI can assist by generating prototype solutions, allowing for iterative testing, and providing analytic tools for benchmarking, enabling students to perform comprehensive evaluations and make informed improvements.

Learning objective

Students can assess a prototype solution generated by an LLM by applying self-developed evaluation criteria.

GenAI role

The LLM serves as a brainstorming tool to generate a prototype solution based on the student-defined problem and parameters. It acts as a rapid prototyping partner, offering an alternative solution that students can analyse and critique.

Lecturer role

The lecturer introduces students to prototyping and benchmarking methods and emphasises critical evaluation. He/she monitors how students use UvA AI Chat, ensuring prototypes are assessed against self-defined criteria rather than accepted at face value.

Learning activities

Students define their problem and create benchmarks. They then generate one or more prototypes using UvA AI Chat. In small groups, they compare the prototypes with their benchmarks, discuss improvements, and iterate with new prompts.

Instructions for implementation

• Ensure students have access to UvA AI Chat.
• Provide students with clear criteria-setting exercises before they use UvA AI Chat.
• Encourage them to document their benchmark list, the generated prototype, and the evaluation process.
• Make space in class for students to share outcomes and reflect on strengths/weaknesses.

Assessment

Students can submit a written report analysing the LLM-generated prototype. The report should include:

• A clear articulation of the problem definition and the rationale behind the chosen benchmarks.
• A description of the LLM-generated prototype.
• A detailed evaluation of the prototype against the benchmarks, including its strengths, weaknesses, and areas for improvement.

Variations

• Students first create their own prototype without UvA AI Chat, then compare it with the LLM-generated version.
• Run a “benchmark fair” where student groups showcase their prototypes and receive peer critique alongside their own evaluations.
• Use UvA AI Chat as a “devil’s advocate” to highlight weaknesses in the prototype rather than as a generator of solutions.

Tips

• Students:
  • Compare the LLM prototype to all your benchmarks and refine it.
• Lecturers:
  • Model how to refine prototypes through iterative prompting.
  • Facilitate discussions about the ethical implications of using LLMs in design and development.
• Reflection question: What are the ethical implications of relying on an LLM for this stage of problem-solving?

Rationale

Analysing a complex negotiation scenario from the perspective allows students to identify key leverage points and potential compromises, enhancing their strategic thinking and negotiation skills. This capability is vital for resolving conflicts and achieving favourable outcomes in diverse professional/academic settings. GenAI can assist by simulating negotiation scenarios, offering role-specific insights, and suggesting strategies, thereby helping students practice and refine their analysis and negotiation abilities.

Learning objective

Students assess a negotiation scenario from a chosen stakeholder’s perspective and identify strategic points of influence and potential areas for agreement.

GenAI role

The LLM acts as the opposing party in the negotiation, responding to student offers and arguments, and offering counter-proposals based on the defined scenario parameters. The LLM can also be prompted to adopt different negotiation styles (e.g., aggressive, collaborative) to add complexity.

Lecturer role

The lecturer frames the negotiation exercise, explaining key principles (e.g., concessions and active listening). They oversee practice sessions, helping students reflect on both content and style of negotiation.

Learning activities

Students are assigned a negotiation case with defined roles. They prepare their strategy, then enter into a negotiation with the GenAI-persona simulating the opposing party. Afterward, students debrief in groups, comparing outcomes and strategies.

Instructions for implementation

• Make sure all students are placed in a single UvA AI Chat group.
• For each stakeholder, develop a persona that acts as the opposing party in a negotiation. Add this persona to the UvA AI Chat group.
• Share example prompts for setting negotiation tone and style.
• Encourage students to log their interactions and annotate them with reflections.

Assessment

Students submit a reflection summarising their negotiation strategy, analysing the LLM’s responses, and evaluating their own performance. The negotiation transcript itself can also be assessed for application of learned concepts. A rubric could be used to assess the quality of the arguments, the identification of leverage points, and the overall outcome of the negotiation.

Variations

• Students rotate roles, first negotiating with the LLM and later with peers, to compare experiences.
• Use escalating difficulty: Start with simple negotiations (price bargaining) and move to complex, multi-stakeholder negotiations.
• Have students replay the same negotiation with the LLM adopting different negotiation styles (e.g., collaborative vs. competitive).

Tips

• Students:
  • Define your goals, priorities, and fallback options before starting the negotiation.
  • Pay attention to how the LLM adapts to your strategies. Don’t just react, but analyse.
• Lecturers:
  • Provide students with clear instructions and parameters for the negotiation scenario.
  • Encourage them to experiment with different approaches and reflect on the LLM’s responses.
  • Consider providing sample dialogues or transcripts to illustrate effective negotiation techniques.
• Reflection question: How did the LLM’s responses influence your negotiation strategy? Did interacting with the GenAI reveal any biases or limitations in its understanding of the scenario?