20Mar’24
This page provides examples for designing AI out of or into your assignments. These ideas are intended to provide a starting point as you consider how assignment design can limit or encourage certain uses of AI to help students learn.
CETLI is available to consult with you about ways you might design AI out of or into your assignments.
Considerations for Crafting Assignments
- Connect to your goals. Assignments support the learning goals of your course, and decisions about how students may or may not use AI should be based on these goals.
- Designing ‘in’ doesn’t mean ‘all in’. Incorporating certain uses of AI into an assignment doesn’t mean you have to allow all uses, especially those that would interfere with learning.
- Communicate to students. Think about how you will explain the assignment’s purpose and benefits to your students, including the rationale for your guidelines on AI use.
Design Out: Limit AI Use
Very few assignments are truly AI-proof, but some designs are more resistant to student AI use than others. Along with designing assignments in ways that deter the use of AI, inform students about your course AI policies regarding what is and is not acceptable as well as the potential consequences for violating these terms.
Assignments that ask students to refer to something highly specific to your course or not easily searchable online will make it difficult for AI to generate a response. Examples include asking students to:
- Summarize, discuss, question, or otherwise respond to content from an in-class activity, a specific part of your lecture, or a classmate’s discussion comments.
- Relate course content to issues that have local context or personal relevance. The more recent and specific the topic, the more poorly AI will perform.
- Respond to visual or multimedia material as part of their assignment. AI has difficulty processing non-text information.
Find opportunities for students to present, discuss their work, and respond to questions from others. To field questions live requires students to demonstrate their understanding of the topic, and the skill of talking succinctly about one’s work and research is valuable for students in many disciplines. You might ask students to:
- Create an “elevator pitch” for a research idea and submit it as a short video, then watch and respond to peers’ ideas.
- Give an in-class presentation with Q&A that supplements submitted written work.
- Meet with you or your TA to discuss their ideas and receive constructive feedback before or after completing the assignment.
This strategy allows students to show how they have thought about the work that they’ve done and places value on their awareness of their learning. For instance, students might:
- Briefly write about a source or approach they considered but decided not to use and why.
- Discuss a personal connection they made to the learning material.
- Submit a reflection on how the knowledge or skills gained from the assignment apply to their professional practice.
Consider asking students to show the stages of their work or submit assignments in phases, so you can review the development of their ideas or work over time. Explaining the value of the thinking students will do in taking on the work themselves can help deter students’ dependence on AI. Additionally, this strategy helps keep students on track so they do not fall behind and feel pressure to use AI inappropriately. For materials handed in with the final product, it can give you a way to refer back to their process. You may ask students to:
- Submit an outline, list of sources, discussion of a single piece of data, explanation of their approach, or first draft before the final product.
- Meet briefly with an instructor to discuss their approach or work in progress.
- Submit the notes they have taken on sources to prepare their paper, presentation, or project.
Prior to beginning an exam or submitting an assignment, you may ask students to confirm that they have followed the policies regarding academic integrity and AI. This can be particularly helpful for an assignment with different AI guidelines than others in your course. You might:
- Ask students to affirm a statement that all submitted work is their own.
- Ask students to confirm their understanding of your generative AI policy at the start of the assignment.
If you decide to limit student’s use of AI in their work:
- Communicate the policy early, often and in a variety of ways.
- Be Transparent. Clearly explain the reasoning behind your decision to limit or exclude the use of AI in the assignment, focusing on how the assignment relates to the course’s learning objectives and how the use of AI limits the intended learning outcomes.
Design In: Encourage AI Use
You may find that assignments that draw upon generative AI can help your students develop the thinking and skills that are valuable in your field. Careful planning is important to ensure that the designed use of AI furthers your objectives and benefits students. Become familiar with the tasks that AI does and does not do well, and explore how careful prompting can influence its output. These examples represent only a small fraction of potential uses and aim to provide a starting point for considering assignments you might adapt for your courses.
Consider using AI tools to generate original content for students to analyze. You might ask students to:
- Compare multiple versions of an AI-generated approach to problem-solving based on the same task.
- Analyze case studies generated by AI.
- Determine and implement strategies for fact-checking AI-generated assertions to examine the value of information sources.
Generative AI may support students as they take on more advanced thinking by offering help and feedback in real time. For instance, students can:
- Input a provided prompt that guides AI to act as a tutor on an assignment. Prompts can lead the AI tutor to review material, answer questions, and help students use problem-solving strategies to find a solution.
- Ask AI to support the writing process by having it review an essay and provide feedback with explicit instructions to help identify weaknesses in an argument. Consider asking students to turn in the transcript of their discussion with the AI as part of the assignment.
- Use AI for coaching (guidance) through complex tasks like helping students without coding skills create code to analyze materials when the learning goal is data analysis rather than coding.
To support students in learning how to test their ideas, to understand what is arguable, or to practice voicing ideas with feedback, generative AI can be prompted to participate in a conversation. You might ask students to:
- Instruct AI to respond as someone unfamiliar with the course material and engage in a dialogue explaining a concept to the AI.
- Find common ground in a discussion of a controversial issue by asking AI to take a counterposition in the debate. The student could question the AI’s contradictions or identify oversimplifications while focusing on defining their own position.
- Engage with AI as it role-plays a persona like a stakeholder in a case study or a patient in a clinical conversation.
Students can engage with AI as a thought partner at the start or end stages of a project, without allowing AI to do everything. The parts of the task AI does and the parts students should do will depend on the type of learning you want them to accomplish. You might ask students to:
- Use AI to draft an initial hypothesis. Then, the student gathers, synthesizes, and cites evidence that supports or refutes the argument. Finally, the student submits the original AI output along with their finished product.
- Brainstorm with AI, using the tool to generate many possible positions or projects. Students decide which one to pursue and why.
- Develop an initial draft of code on their own, ask AI for assistance to revise or debug, and evaluate the effectiveness of AI in improving the final product.
The specific prompts a user gives to an AI tool strongly determine the quality of its output. Learning to write effective prompts can help students use AI to its fullest potential. Students might:
- Prompt AI to generate responses on a topic the student knows a lot about and evaluate how different prompt characteristics impact the quality and accuracy of AI responses.
- Research applications of prompt engineering that are emerging in their field or discipline.
- Create custom instructions for AI to help with a specific, challenging task. Share their results with peers and evaluate the effectiveness of each others’ results.
If you allow students to utilize AI in their work:
- Make it clear that students are responsible for any inaccuracies in content or citations generated by AI, and that they should own whatever positions they take in submissions.
- Set and communicate clear expectations for when and how AI contributions must be acknowledged or cited.
- Consider barriers that prevent equitable access to high-quality AI tools. Students may also have varying levels of AI knowledge and experience using the tools. CETLI can assist you in developing an inclusive plan to ensure students can complete the assignment.
If students may use AI with proper attribution, APA, MLA, and Chicago styles each offer recommendations for how to cite AI-generated contributions.
Harvard’s metaLAB AI Pedagogy Project provides additional sample assignments designed to incorporate AI, which are free to use, share, or adapt with appropriate attribution.
CETLI Can Help
CETLI staff are available to discuss ideas or concerns related to generative AI and your teaching, and we can work with your program or department to facilitate conversations about this technology. Contact CETLI to learn more.