How to Incorporate Generative AI

A concept map is a visual representation of information and relationships. Concept maps can take the form of diagrams, flow charts, tables, and more.

Supporting Integrity

• Because concept maps combine multiple modes of information through diagrams, drawings, and text, they are excellent ways for students to show their thinking process but are too multimodal for AI to complete well.
• Concept maps, diagrams, and notes can be included as assignment components. Multiplying the ways students are asked to represent their learning can diminish the usefulness of academically dishonest AI use as well as support more authentic assessment of learning.
• Concept maps can be assigned after an assignment has been completed as a means for students to represent the knowledge and understanding they realized through the assignment.
• Ask students to develop a concept map over time, adding to it or revising it at the end of each week or unit, or at midterm and end of term based on new course content students read, watched, listened to, or thought about. Students can also be given other students’ concept maps to add to or revise. These activities can be done individually, in pairs, or in groups.

Incorporating Generative AI

• Ask generative AI to summarize a large concept or course topic in writing. Students can then be challenged to interrogate the output as they create concept maps that that are robust, relevant to the course, inclusive, and equitable by, for example: making decisions about what content to include and exclude, filling in gaps, making connections, adding citations, and addressing biases.
• Have students use a generative AI output such as an AI-generated Infinite Conversation to create a concept maps that shows connection and/or plots fundamental arguments, as well as reveals gaps, exposes biases, and shows superficiality. Students can then be tasked with making improvements to the original concept map through further research or by integrating their course-based learning.

A content summary provides a synopsis and the most salient or pertinent components of a journal article, book, film, or other source material.

Supporting Integrity

• Students can be asked to include specific types of connections in their content summaries, for example, connections between the source material and specific course content, one or more specific course materials, other courses they’ve taken, personal experiences, a pop-culture reference, a specific artifact (image, song, artifact), etc.
• Request that content summaries speak to the source material’s relevance to the course, a specific unit or theme of the course, a specific assignment or research question, etc.
• Students can be asked to comment on source materials’ credibility, currency, objectivity, etc.

Incorporating Generative AI

• Have generative AI summarize the last course essay or another piece of course material. Ask students to point out the shortcomings and missed nuance of the AI output and then have them reflect on how nuance shapes their overall understanding of a topic.
• Have generative AI summarize several pieces of course or essay content and build a wider literature review. Consider prompts such as: What themes has the generative AI tool focused on? What themes has it ignored? What do its biases and gaps appear to be and why are those important for understanding the topic?

An experiment is a procedure or other activity carried out to make a discovery, test a hypothesis, illustrate a known fact, or practice skills.

Supporting Integrity

• Generative AI can’t be of assistance if its use is prohibited, especially during an experiment that is directly observed.
• Generative AI is of limited assistance during experiments that require physical manipulations or interpersonal interactions.

Incorporating Generative AI

• Students can use generative AI to identify materials and/or steps necessary to carry out an experiment or to augment opportunities for success.
• Students can consult generative AI after a failed experiment to troubleshoot the ‘why’ of the failure.

Infographics are visual representations of ideas, information, and data.

Supporting Integrity

• Focus grading rubrics on how concepts are illustrated and connections between information rather than the artistic product, which can be assisted, or components designed, by generative AI.

Incorporating Generative AI

• Have a generative AI tool output a base graphic for students to augment, correct, and build out based on set criteria.
• Explore having generative AI create images for infographics. PowerPoint’s “Slide Designer” is a generative AI tool that’s readily available for students to experiment with. It creates slide designs based on slide content.

A timeline is a visual representation of a chronological sequence. Students can either create their own timelines or critique and edit sourced timelines.

Supporting Integrity

• Ask students to include specific types of connections in their timelines to diminish meaningful assistance from generative AI. For example, ask students to draw connections with other course materials, specific course materials, other courses, personal experiences, a pop-culture reference, a specific artifact (image, song, artifact), etc.

Incorporating Generative AI

• Use generative AI to produce a timeline or to identify the precipitators of a historical events (e.g. “Summarize the causes of the French Revolution”) and have students use this output  to consider how narrative and perspective are subject to bias, how generative AI and the historical cannon privilege some stories while erasing others, etc.
• This example of an artist who harnesses AI to “simulate time travel” selfies is an example of how someone can tell AI how to construct elements of a scene that are inspired and informed by course content.

Active learning is an approach to instruction that engages students with course content through discussions, debates, simulations, games, role playing, problem solving, one minute papers, etc. (explore 226 active learning techniques). In this way, responsibility for learning and metacognitive skill development is shifted toward the learner and away from the instructor. Active learning can take place in the classroom, online, and outside of the classroom and can be assessed as a part of participation grades or through submitted learning artifacts in engagement portfolios collected through MyLS DropBox throughout the term.

Supporting Integrity

• When active learning takes place in synchronous in-person or online settings, the opportunities for generative AI to provide meaningful and timely assistance is diminished.
• Choose and create activities that permit active learning time to be technology-free time.
• When active learning takes place in any setting, have students engage in activities that encourage making deep connections, referencing student-generated content (e.g. discussion board posts, student presentation content, peer feedback, etc.) and incorporating course-specific content (e.g. lecture materials, guest lecture content, field trips, previous in-class activities, etc.)
• Require work associated with learning activities (e.g. discussion notes, debate outline, problem proofs/solutions, personal reflection on the activity, etc.) to be included in a portfolio diminishes the extent to which generative AI can be productively used.
• Focus active learning on very recent events, discoveries, publications, etc. Generative AI output is only as current as the dataset the tool was trained on thus limiting its ability to process questions related to current events.

Incorporating Generative AI

• Students can compete to develop the most meaningful, complete, inclusive, or course relevant definitions of course concepts with half the class using generative AI and half of the class using other tools and resources
• Have generative AI produce a case study, scenario, problem, poem, image, song, etc. for students to analyze, comment on, interrogate, or solve. To add complexity, students can be asked to do this from a particular framework, point of view, era, etc. or using instructor-created prompts.

A bibliography is an alphabetized list of sources, such as books, journal articles, websites, newspaper databases and other documents that are relevant to a course or project. An annotated bibliography provides a brief (approximately 150 word) summary and assessment of each source.

Supporting Integrity

• While generative AI can produce relevant and well-written annotated bibliographies, these are unlikely to include very recent sources (sources published in the last 12-24 months) and its assessments of sources may not be course- or assignment-specific, speak to the relevance of the article to the student’s research topic, or address the reliability of the sources.

Incorporating Generative AI

• Have students design a question and ask generative AI to produce an output (e.g. essay, article, op-ed, etc.) based on an annotated bibliography that the student created or that they asked generative AI to create. Next, have students take the output and review it for accuracy and completeness before correcting and adding both content and citations to improve upon the generative AI output. Consider adding an assignment component that asks students to reflect on what they learned during this process.
• Have AI generate an annotated bibliography for a topic that is course-adjacent and have students critique what’s generated using what they’ve learned in the course to date.

A case study is an in-depth, detailed, and multi-faceted presentation and/or examination of a particular event, scenario, person, group, or organization within a real-world context.

Supporting Integrity

• While generative AI tools like ChatGPT can write and analyze cases in ways that are stylistically in-line with case writing style, its outputs lack depth and its ability to find novel solutions to complex problems is limited.
• Use case studies focused on very recent events (i.e. no more than 18 months old). Generative AI output is only as current as the data set it was trained on thus limiting its ability write and analyze very current cases.
• Have students analyze an imaginary case study. Incorporating fictional or fictitious characters, imaginary places, and fantastical events into a case study can frustrate the ability of generative AI to produce meaningful output.

Incorporating Generative AI

• Have generative AI write a case study and have students correct it, expand upon it,  consider it from gendered lenses or Indigenous perspectives make it more inclusive, add citations, add relevant academic references, or analyze it in one or more course-specific manners. For example, using ChatGPT, begin your query with “write a case study about” to get output in a more traditional case writing style (e.g. “write a case study about a small food service business trying to attract new customers with limited marketing budget”).
• After students write their own case studies, have generative AI solve them using different input instructions to help students see their cases from different perspectives.
• Have generative AI summarize sections of student-written case studies to help students see salient points more easily as they build their executive summaries.

Students read an instructor-selected text and respond to questions focused on specific content, skills, and learning outcomes.

Supporting Integrity

• Create prompts that focus on making connections with personal experiences (e.g. How does this relate to something you learned or experienced in a previous week’s class? Outside of class? In another course? When you were younger?) and with student-generated content (e.g. How does this relate to a class discussion, discussion board post, student presentation, etc.?)
• Have student engage in social reading via whole class or small group book clubs or literature circles. Social reading gives students opportunities to design their own learning, gain confidence in discussions, analyze and critique their peers’ arguments, consider perspectives different from their own, and develop social skills and connections.
• Have students engage in social annotation. Social annotation requires learners to comment on assigned text with or without reading prompts, review and engage with the comments of others, connect the assigned text with prior knowledge, and position the assigned text within the context of personal experience.

Incorporating Generative AI

• Have generative AI produce a summary of a chosen text and have students interrogate it for bias, relevancy to the course, authenticity, etc.
• Have students compare their answers to prompts with generative AI’s answers to the same prompts and then consider questions such as: What are the similarities and differences? How do you feel about generative AI’s answers? To what extent do self-authored answers versus generative AI’s answers impact your knowledge, understanding, skills, and progress toward learning outcomes? This can be used as a synchronous/in-class learning activity, as a graded or ungraded assessment, or as a learning artifact for inclusion in a portfolio.

Students work in pairs, groups, or teams to produce a specific end product such as an essay, report, case study, prototype, presentation, poster, etc.

Supporting Integrity

• Require students to reference the sources used to produce the end product. Language model generative AI mines a huge sample of text taken from the internet and uses it to generate outputs. Requiring both academic and non-academic references from a combination of specified and students’ choice materials can discourage students from submitting AI-generated assignments.
• Require students to engage with feedback. Build an instructor, TA, or peer review process into assignments and require students to make revisions or otherwise respond to feedback in a staged submission process.

Incorporating Generative AI

• Use generative AI as collaborator. Have students and their AI collaborator give each other feedback on their writing. This can help students reflect on how their writing style is unique, improve their ability to communicate what they want to say, identify biases in generative AI outputs, and find an authentic voice in their writing. Students can be asked to produce a reflection on the positive and negative impacts generative AI had on their ability to communicate effectively and authentically.
• Juxtapose generative AI and human collaboration. Have pairs of students, each student with their own research question, collaborate with a generative AI tool and each other to produce a final product. Students work together to coach a generative AI tool for assistance with each research question and also give feedback to each other at various points through the assignment process. Students can be asked to use track changes to show how their essay (or other final product) was improved along the way. After assignment submission, students can be asked to reflect on the process of working with the generative AI tool and their student partner with prompts such as: What worked well and what didn’t? How did your student partner help you produce a better final product? How did the generative AI tool help you produce a better final product? What were the benefits and drawback of your student partner versus your generative AI partner?
• Generative AI can be used to suggest effective wording or alternate phrasing for written assessments.

Exams and tests assess knowledge, proficiency, or skill in at a particular point in time.

Supporting Integrity

• Write questions that include student-generated content (e.g. discussion board posts, student presentation content, peer feedback, etc.) or course-specific content (e.g. lecture materials, guest lecture content, field trips, experiments and other in-class activities, etc.).
• Create questions that focus on making connections with personal experiences (e.g. How does this relate to something you learned or experienced in a previous week’s class? Outside of class? In another course? When you were younger?) and with student-generated content (e.g. How does this relate to a class discussion? Discussion board posts? A student presentation? etc.)
• Build in metacognitive pieces that ask students to reflect on the process of learning, how they arrived at their answers, and what techniques they used to study and grow their  knowledge.
• Ask students to include specific types of connections in their answers. For example, ask them to draw connections between specific course materials, personal experiences, a pop-culture reference, a specific artifact (image, song, artifact), etc.
• Require students to include student-generated and course-specific content in their responses.
• Require students to submit graphs, calculations, diagrams, etc. that must be produced by the student in order to arrive at an answer.
• Use very current topics, events, issues, publications, etc. in questions or require their reference in answers. Generative AI data sets typically lag by about 12-18 months.
• Consider incorporating low-stakes, multiple-attempt, and two-stage testing into your teaching practice. De-centering grades can decrease the temptation to cheat, encourage self-evaluating knowledge, understanding and skills, and identify learning gaps.
• Have students write some exams and tests in person.

Incorporating Generative AI

• If the test or exam is “open book” require students to cite the resources they use, including generative AI.  Find more strategies for take home open book” exams later in this resource.

An experiment is a procedure or other activity carried out to make a discovery, test a hypothesis, illustrate a known fact, or practice skills.

Supporting Integrity

• Generative AI can’t be of assistance if its use is prohibited, especially during an experiment that is directly observed.
• Generative AI is of limited assistance during experiments that require physical manipulations or interpersonal interactions.

Incorporating Generative AI

• Students can use generative AI to identify materials and/or steps necessary to carry out an experiment or to augment opportunities for success.
• Students can consult generative AI after a failed experiment to troubleshoot the ‘why’ of the failure.

Fact Sheets ask students to identify and communicate relevant information or evidence to illuminate or frame a particular issue, problem, need, event, or subject. Policy Briefs typically include the additional step of making a recommendation or a rank-ordered set of recommendations to a policymaker.

Supporting Integrity

• Ask students to produce fact sheets or policy briefs on very current topics. Generative AI data sets typically lag by about 12-18 months.

Incorporating Generative AI

• Generative AI is good at summarizing and can be employed to identify salient information to include in a fact sheet or policy brief.
• Students can be asked to be fact-check and verify information gleaned from generative AI.
• When working on a policy brief, students can “compete” with generative AI to produce the most persuasive and compelling recommendation.

A structured written response to a complex or probing research question(s).

Supporting Integrity

• Focus on the process of writing rather than the final written output and have students submit research questions, drafts, outlines, and references along with the finished essay. These can be submitted for review as set points during the course of the project or collected with the final essay submission.
• Ask students to include specific types of connections in their essays. For example, ask them to draw connections with other course materials, specific course materials, other courses, personal experiences, a pop-culture reference, a specific artifact (image, song, artifact), etc.
• Scaffold essay components, have students respond to peer and instructor feedback at each stage or at specific points in the project, and add a reflective component that asks students to reflect on how feedback contributed to producing a better final product.

Incorporating Generative AI

• If use of generative AI is permitted, require students to both cite it in the body of the essay and include all of its raw outputs based on their queries as appendices.
• Generative AI can be used to suggest effective wording or alternate phrasing as a means of improving the effective communication of ideas and information
• As above, scaffold essay components but incorporate generative AI tools and feedback alongside instructor and peer feedback. Students can map the development of their essay through drafts and write a concluding reflection on the roles of both generative AI and peer or instructor feedback in their essay development process.

A literature review is a comprehensive summary of scholarly work on a particular topic, concept, or theory.

Supporting Integrity

• Ask students to include an assessment of relevancy (e.g. to course topics, course themes, student discussion board posts, etc.) for each source in their literature. While generative AI can write summaries and broad overviews of common topics, it’s not able to assess relevance.
• Have students assess the authority, reliability, and currency of each source, something generative AI is not well suited to do.
• Ask students to include specific types of connections in their literature reviews. For example, ask students to draw connections with course topics, specific course materials, personal experiences, a reference, etc.

Incorporating Generative AI

• Ask generative AI to produce a literature review on a particular topic, concept or theory. Have students assess the literature review for bias, currency, omissions, etc. and then revise it.
• As an extension of the above, have students work on the same topic, concept and theory in pairs or groups and then have them compare the results of their generative AI literature reviews, assessments, and improvements. Ask students to reflect on which group’s final product was the best and why.

In an observational assessment evaluative information about knowledge, skill, and ability is obtained through direct observation by an evaluator (e.g. instructor, TA, lab assistant, peer, or self.)

Supporting Integrity

• Generative AI cannot provide assistance if its use isn’t permitted during the assessment.
• Students can be asked to submit a personal reflection on their preparation for the assessment, what helped or hindered their performance during the assessment, what they would do differently if they could repeat the assessment, and what steps they can take to improve their performance going forward.

Incorporating Generative AI

• Students can use generative AI as a preparatory tool to advance knowledge and hone skill before they engage in an observational assessment. For example, they can ask generative AI to summarize the steps necessary to complete a particular task or to identify common task errors as a means of studying.

In an open book “take home” exam or test students have permission to access course materials and other aids while completing the assessment. Questions require students to apply, analyze, and evaluate course materials in creating their responses.

Supporting Integrity

• Write questions that include student-generated content (e.g. discussion board posts, student presentation content, peer feedback, etc.) or course-specific content (e.g. lecture materials, guest lecture content, field trips, previous in-class activities, etc.).
• Require students to include student-generated and course-specific content in their responses.

Incorporating Generative AI

• Have students use generative AI to create a topic outline or bulleted skeleton of an answer and then improve upon it with course-based learning.
• If use of generative AI is permitted, have students cite generative AI contributions, submit the queries they made to generative AI along with the raw outputs that resulted, and provide a reflection on their use of generative AI.

During peer evaluations students review, assess, and provide feedback on other students’ work.

Supporting Integrity

• Have students write a reflection on the peer feedback received using prompts such as: What do I agree with? What do I disagree with? What will I do (or have I done) to address the feedback? How will (or did) the feedback I received make my next draft (or final product) better?
• Have students submit a “response to peer feedback” statement with next or final submission explaining how they addressed the feedback they received or why they chose to reject it.

Incorporating Generative AI

• Require students to use generative AI as a peer, offer feedback on its research and writing, and suggest avenues for further exploration that reveal richer connections and application of information.
• Have students design a question and have generative AI write an article based on that question. Next, have students engage in a scholarly review of the article generated including comments, corrections, suggestions for improvement, etc.

A portfolio is a collection of artifacts of learning (e.g. reflections, annotations, discussion board posts, assignments, test answers, etc.) Students engage in a variety of learning activities throughout the term and then compile a portfolio to show evidence of their learning journey.

Supporting Integrity

• Requiring a wide variety of learning activities to be included in a portfolio diminishes the extent to which generative AI can be productively used.
• Asking students to reflect on their portfolio content and their growth throughout the course helps enrich these collected artifacts of learning while at the same time being an assessment that generative AI cannot complete.
• Students can be asked to annotate earlier work in their portfolio and draw connections between assessments, course themes, and skills in a concept mapping style as additional means of limiting the productive assistance of generative AI.

Incorporating Generative AI

• Generative AI can be a collaborator or contributor to individual portfolio items and its contributions can be cited.
• Generative AI can give feedback on portfolio items like written work and provide comments, feedback, and next steps that students can pursue in their research. Students can integrate or ignore the feedback, provide rationale for these decisions, and reflect on other ways to advance their knowledge, understanding, skills, etc.

A poster presentation is a visual representation of student research, typically demonstrating knowledge of theory, literature review, methods, and findings.

Supporting Integrity

• While generative AI can produce poster elements, layout ideas, executive summaries and graphics for poster presentations, the multimodal nature of poster presentations requires students to make and reveal deep connections and organize their thinking.

Incorporating Generative AI

• Generative AI is useful for summarizing key points, drafting executive summaries, and prioritizing information for inclusion on posters.
• Generative AI can help design graphics and suggest layouts for posters, assisting with the organizational flow of the end produce. For example:
  • Image generators can produce images that can help illustrate concepts or provide visual excitement in a poster;
  • PowerPoint’s Slide Designer feature uses generative AI to suggest layouts based on content. Using it as an idea generator can help students see their information presented in different ways instantly.

Students explain a concept, process, idea, project, experiment, etc. to others orally or audio-visually. Presentations can be delivered live (in person or virtually) or recorded.

Supporting Integrity

• Presentations are excellent ways to augment other written assignments to extend learning and using multiple means to assess student learning holistically.

Incorporating Generative AI

• Generative AI is especially useful for summarizing key points, which can be included on slides.
• Generative AI can help design graphics and slides for presentations.
  • Image generators can produce images that can help illustrate concepts or provide visual excitement in a presentation.
  • PowerPoint’s Slide Designer feature uses generative AI to suggest layouts for each slide based on its content.

Supporting Integrity

• While generative AI could be used to assist with various aspects of a prototype project, its ability to be creative is limited. The multifaceted nature of prototyping and testing leads to scaffolded assessments with clear draft and revision stages. Documenting the process by which prototypes were created is inherent in prototyping, and assembling those documents emphasizes the process of creation and revision rather than the final product.

Incorporating Generative AI

• Generative AI, like ChatGPT, can debug and write code as well as natural language text. Generative AI can become a debugging tool used to help diagnose problems with prototypes and code.
• Generative AI, like ChatGPT, can be used to summarize papers and other text to help identify what is most salient in the writing. This can help improve written communication and guide where to place emphasis in grant proposals, pitches, etc.

Reflection papers ask students to consider what they have learned in the context of their lived experiences, use what they have learned to inform future action, or consider the real-life implications of their thinking.

Supporting Integrity

• Generative AI can’t connect to meaning, personal experience, or feeling. Reflection and connection with personal experiences is a key way to not only reduce the impact of generative AI tools on academic integrity, but also for students to make meaning from their learning.

Incorporating Generative AI

• Ask generative AI “if-then” questions and have students offer personal reflections on the output or interrogate the output for biases, gaps, and other shortcomings.

A scaffolded assessment is a larger project, case, problem, or assignment broken up into smaller, progressive learning activities (“chunks”) that build toward a final summative assessment.

Supporting Integrity

• Generative AI doesn’t incorporate reflections, make meaningful connections, draw complex conclusions, or handle iterative feedback well. Having students build a scaffolded assessment through brainstorming, outlining, researching, iterative research or literature summaries, and responding to feedback can minimize unauthorized use of generative AI.
• While generative AI can potentially be used to produce pieces of a scaffolded assignment (such as draft summaries of research sources), it cannot create the varied and iterative steps that lead to a final scaffolded product.

Incorporating Generative AI

• Students can be permitted to use generative AI to produce pieces of a scaffolded assignment (e.g. summaries of research sources, a procedural outline, a timeline, etc.) or as an approved source for basic information about a topic, event, person, procure, etc. Remember to require students to cite generative AI.
• Generative AI can be used to suggest effective wording or alternate phrasing for written work.
• Generative AI can be used to check drafts for gaps, problems, and other deficiencies.

The Three Minute Thesis (3MT) is a popular competition among research-based graduate students that can also be used as an assessment tool for all students. In a 3MT students are allowed one slide and three minutes to present course content, a specific topic, research, or a project to a general audience.

Supporting Integrity

• Presentations are excellent demonstrations of learning, and the act of summarizing and choosing what to highlight stimulates deep learning. While generative AI can summarize information and provide broad insights into what themes emerge from theses and research papers, they lack the personality and creativity to produce dynamic presentations.
• Students can be asked to include specific types of connections in their Three-Minute Thesis presentations. For example, connections with other course materials, specific course materials, other courses, personal experiences, a pop-culture reference, a specific artifact (image, song, artifact), etc.

Incorporating Generative AI

• Ask students to paste sections or the entirety of their thesis into a generative AI language model like ChatGPT with the precursor “Summarize:” or “Summarize this:” and have them reflect on the results. Have students regenerate their responses repeatedly to see how the nuance of the output changes and gain new perspectives about what stands out or might be veiled. Have students write a reflection or make a presentation to the class that discusses their findings.
• Generative AI can help design graphics and slides for presentations. For example, Image generators can produce images to help illustrate concepts or provide visual excitement in a presentation, and PowerPoint’s Slide Designer tool uses generative AI to suggest layouts for each slide based on the slide’s content.