The Final Assignment, Reimagined

Dear Colleagues,

Welcome to the 42nd edition of Weekend Reads from CEPI.

  Scene One: A Familiar Classroom

Picture the final week of the semester. Students have spent weeks on their projects, and one group after another comes to the front of the room. The slides are polished, the transitions are smooth, and the content is neatly organized. Some presentations look almost professional.

Then someone asks a question. “Why did you choose this approach?”

One student hesitates. Another repeats a line from the slide. A third gives a confident answer that sounds right and comes apart at the second follow-up question.

Perhaps the students understand more than they can say in that moment. Perhaps they learned a great deal. Perhaps the polished deck covered real gaps in their understanding. Often we cannot tell which.

Generative AI has made this harder to ignore. A student can now research a topic, outline it, build a polished deck and write speaking notes with very little effort. None of those steps needs the depth of understanding the assignment was meant to build.

Presentations can still be good assessments, especially when students must defend their reasoning and answer questions they did not see coming. If the finished deck is our main evidence of learning, though, we may be measuring something we never meant to measure.

  Scene Two: Another Classroom

Imagine another class in the same final week. These students have also spent weeks on real-world problems. They have gathered evidence, read data, interviewed people and built a growing understanding of the issue in front of them.

Here, no team walks to the front to present a deck. Each team designs and leads a thirty-minute learning session for their classmates, and the questions that unsettled the first room become part of the plan.

In this assignment, students are expected to use AI throughout. Over the course, they have learned to strengthen their research and design work with AI tools. Their own judgment and decisions stay at the center of that work. They can use AI to research, organize information, draft the packet, design activities, build slides and prepare their facilitation notes.

The packet must be accurate, readable, properly sourced and worth a classmate’s time. Students answer for every claim and every sentence in it, however it was produced.

The workshop must be designed with care. Every activity needs a learning purpose. The team has to anticipate misunderstandings, handle unexpected questions, manage participation and adjust when the room heads somewhere new. Each student shows their own understanding through how they facilitate, question, explain and reflect.

Teams may also invite a few guests with a real connection to the problem. A guest might be a faculty member from another discipline, a staff colleague, a family member or someone from the wider community. Guests bring perspectives the class does not have, and they make the workshop part of the SFBU community.

Checkpoints run through the weeks before the final, for the instructor and for peers. Each team first submits a proposal naming the idea it will teach. The instructor and another team then review a draft of the reading packet. A short run-through with the instructor tests the activity before the day. During the workshop, classmates’ questions test each facilitator. Afterward, each student answers one question alone and writes a short reflection.

A strong team workshop can still carry a member who has not mastered the material. For that reason, every checkpoint looks at individual students as well as the team.

Thirty minutes is deliberate. It gives the students leading the workshop time to show a real grasp of the subject. It also gives participants time to learn something they can show. With classes capped at twenty, every team has room for a full session. Faculty can choose a different length for their own context, as long as it leaves room for both.

  Scene Three: Your Classroom

The second classroom may look different from yours. You may teach accounting, mathematics, computer science, business or English. Your final may be an exam, a paper, a case analysis, a presentation or a technical project. A workshop will suit some of those and not others.

Now that AI can complete large parts of many tasks, it is worth looking again at what your assessment asks students to do. Some assessments need a small adjustment. Others need a fuller redesign. Some already give clear evidence of learning, and finding that out is useful too.

Across higher education, many of us are returning to handwritten exams and other familiar safeguards. It is a reasonable response when we have no other approach we know well and trust to work. It does not yet answer the question AI has raised about what students should learn and show.

Before returning to those safeguards, work through these four questions.

If your syllabus leaves room to change the final this term, tell students early and in writing. If it does not, plan the redesign for next term.

Wishing you a reflective and restorative weekend.

Warmly,

Center for Empowerment and Pedagogical Innovation (CEPI)

© 2026 CEPI. All rights reserved.