Global Assignment Trends

Full disk view of the Earth taken on December 7, 1972, by the crew of the Apollo 17 spacecraft en route to the Moon at a distance of about 29,000 kilometres (18,000 mi). It shows Africa, Antarctica, and the Arabian Peninsula.
A high-level view: “The Blue Marble”

Global insights emerging in generative AI research suggest that we should prioritize assignments that are…

  • Transparent on the role of conversational agents
  • Interesting to students
  • Process-focused

Further, we should re-align our assessment criteria to de-emphasize traditional at-home writing as the determining factor of a final grade. This may mean adding elements that are…

  • Oral
  • Team-based
  • Demonstrative of contextual knowledge

Think through the consequences of your changing assessment criteria carefully and communicate the new criteria clearly to students from the beginning of the semester. Consider how changes might impact students with disabilities or non-native English speakers. For instance, while a return to in-class proctored pen-and-paper exams will certainly decrease student reliance upon ChatGPT, it may also disadvantage non-neurotypical students and educationally disadvantaged students. Similarly, replacing part of the final paper’s grade with an in-person office hours visit to discuss the ideas behind the paper will help you decide if the student is meeting the class goals around using writing to develop critical thinking. However, it may also open new possibilities for implicit bias.

There is no neutral status quo – teaching as if it were still a pre-ChatGPT world will have its own educationally disadvantaging effects.  But it is important to ask: “are the tradeoffs worth the adaptation?”

Transparent

A good assessment strategy will include generative AI guidelines that are clear about when students can and can’t use conversational agents like ChatGPT. Do not assume that you and your students share the same expectations and norms.

Showing students how to document their interactions with generative AI can alleviate your concerns about academic integrity while also providing a valuable workplace skill. Make sure that they know how you would like them to cite AI-derived content. Ask them to keep notes on the intersection of their ideas and conversational output. Consider requiring chat logs be submitted as a process document. 

Providing a template for acknowledging the use of conversational agents in their assignments will help students know what you are looking for. Western Michigan’s “AI in the Syllabus” resource provides an example. 

AI Acknowledgement: For every assignment submission, you will include a 150-300 word acknowledgment. Your acknowledgement should include, 1. The identification of the tool(s) you used, 2. An explanation of why you decided to use the tool(s), 3. A description of how you used the tool(s) to manage assignment requirements, and 4. A reflection on your experience using the tool, exploring what worked or didn’t, and acknowledging limitations of the tool for this assignment, potential biases, etc. If you opt not to use AI tools, when not required, please use the “AI Acknowledgement” to highlight your non-AI approach and/or your reasons for deciding not to use certain tools.

Interesting to students

Write assignments that make students want to do them, either because the topic is inherently satisfying to them (intrinsic motivation) or application to some external goal is immediately apparent (extrinsic motivation). 

Allowing more student autonomy in topic selection can lead to higher intrinsic motivation and can boost an assignment’s inclusivity. For instance, assigning a topic or project (e.g., “write an infographic shows how to recognize drowning and prevent drowning deaths for lifeguards in training”) may be less interesting to students with no interest in swimming. Instead, ask students to come up with a resource based on what they know of a hobby or professional community’s needs. The gap analysis becomes as important a part of the assignment as the end product, which shifts emphasis away from rote answers and toward creative invention and audience analysis. 

Grades are a famous example of extrinsic motivation, but for many students alignment with expected professional demands is even more motivating. Assigning workplace genres taken from an employer many of your students hope to work for will be more interesting to them than a generic version of that same document. For example, if your institution has a strong Amazon pipeline, consider assigning six-pagers instead of a generic business memo or proposal. 

Oral

While there is a growing consensus that it is a dead end to “out-prompt” ChatGPT in assignment design, changing some assignment formats to emphasize oral communication may be a way to gauge student learning. There are several popular assignment formats already in use that could be adapted for your class, including:

Verbal presentation: instead of asking students to write a report on a topic, consider asking them to present findings to the class and incorporate a substantive Q&A section. This will require some preparation on the instructor’s end to be sure that relevant and challenging questions are posed to the student. However, it may also help your students with real-world environments that are high-stakes and time-bound, such as interviews or final internship presentations.  

Required office hours visit: talk about each student’s ideas with them one-on-one as part of their writing process. Have them explain what their argument is, defend their methodological choices, and explain how they will find and use relevant evidence to support their argument. Take notes on these visits so that you remember what was discussed when you see their written work. 

In-class group meetings: assign a group project that involves in-class meetings in which individual students take leadership roles. Then, circulate around the small groups during class time and sit in on as many of these meetings as you can. Assess each student on how well they facilitate the meeting. This should be done in tandem with a written assignment in which students and their groupmates note their own assessment of the meeting’s effectiveness.

Team Based

Team-based assignments are notoriously difficult to assess and often perceived by students as unfair, so incorporating them into your class already takes a good deal of thought and care. Many of the steps that an instructor might already take to ensure fairness among an all-human team (defining tasks and roles ahead of time, oral and written reflections and peer evaluations, multiple submission milestones, etc.) can also be valuable in a hybrid human-AI team.

Consider casting ChatGPT as an extra (junior) team member. Have students reflect upon their own strengths and determine how they can contribute best to the team. Also have them consider ChatGPT’s strengths and create a similar plan for its contributions. 

Demonstrative of contextual knowledge

Consider revisiting your pre-ChatGPT rubrics to align your assessment criteria more closely with what you see as an assignment’s most creative outcomes. For instance, instead of placing a heavy weight upon an annotated bibliography, challenge students to identify gaps in current literature and to propose their own solutions. Ask them what their research question is and why they chose their methodology to explore that question. They should be able to reflect both on what is there and what is not there in current literature. Summary and description of current literature may be aided by ChatGPT, but students should see this as merely the beginning. 

References

AI in the Syllabus,” Western Michigan University, 2023

Chang, Yunjeong and Peggy Brickman, “When Group Work Doesn’t Work: Insights from Students,” CBE Life Sci Educ. 17(3): 2018.

Evans, Miriam and Alyssa R. Boucher. “Optimizing the Power of Choice: Supporting Student Autonomy to Foster Motivation and Engagement in Learning,” Mind, Brain, and Education. 9(20). 2015: 87-91

Forshell, Johan and Karin Forslund Frykedal, and Eva Hammar Chiriac, “Teachers’ perceived challenges in group work assessment,” ed. Sammy King Fai. Cogent Education 8(2), 2021.

Gimpel, Henner and Kristina Hall, Stefan Decker, Torsten Eymann, Luis Lämmermann, Alexander Maedche, Maximilian Röglinger, Caroline Ruinder, Mareike Schoop, Steffen Vandirk, Manfred Schoch, and Nils Urbach. “Unlocking the Power of Generative AI Models and Systems such as GPT-4 and ChatGPT for Higher Education: A Guide for Students and Lecturers,Hohenheim Discussion Papers in Business, Economics and Social Sciences No. 02-2023, Universität Hohenheim, Fakultät Wirtschafts- und Sozialwissenschaften, Stuttgart. 2023.

How do I cite generative AI in MLA style?MLA Style Center, 2023.

Price, Margaret. Mad At School: Rhetorics of Mental Disability and Academic LifeUniversity of Michigan Press, 2011. 

Stephanie Pulford, Jiahui Tan, Michael Raymond Gonzalez, and Amanda Modell. “Satisfaction: Intrinsic and Extrinsic Motivation in Engineering Writing Coursework“. 2018 ASEE Annual Conference & Exposition , Salt Lake City, Utah, 2018, June. ASEE Conferences, 2018. https://peer.asee.org/30949 Internet. 20 Aug, 2023

Tai, Joanna and  Paige Mahoney, Rola Ajjawi, Margaret Bearman, Joanne Dargusch, Mary Dracup & Lois Harris, “How are examinations inclusive for students with disabilities in higher education? A sociomaterial analysis,” Assessment and Evaluation in Higher Education, 43(3). 2023: 390-402, DOI: 10.1080/02602938.2022.2077910

UNSW Sydney Teaching, “Assessing by Group Work.”