by Zheng Song
I regularly teach CIS 427, Computer Networks. Over the past five years, I’ve taught this course eight times, missing only two semesters. Each time, I’ve tried to make small adjustments to see how students respond and to decide whether the changes are worth keeping for the next round.
My first major change was introducing PBL (Project-Based Learning), asking students to build an IoT system implementing MQTT, a network protocol commonly used in IoT environments. Many students found it exciting because they could get their hands dirty and build something tangible. However, they also found it time-consuming, as most were unfamiliar with IoT concepts and had to spend a lot of time searching for online resources. Some example implementations can be found here: Door Lock Project. Considering that much of their effort went into learning IoT rather than networking, I eventually decided to remove that component in later versions of the course.
My second change involved introducing Kahoot, a live, game-based learning platform that allows students to answer quiz questions in real time using their phones or laptops. I divided the class into teams that competed by answering questions drawn from the lecture material. The instant feedback and leaderboard system created a lively, competitive atmosphere in the classroom. Students discussed their reasoning with teammates, and the room often filled with laughter and energy. This activity encouraged students to recall and apply what they had learned from my lectures, rather than simply memorizing facts. I found that it not only reinforced key concepts but also helped build community and enthusiasm for the subject.
After 2022, as Generative AI (GenAI) tools became widely used by students, I began asking myself a serious question: if students can easily find all my course materials online and use GenAI to solve their problems, why should they still take my class instead of learning on their own? I enrolled in a seven-week course on GenAI in education and learned a great deal—how GenAI can help improve course design, why we should clearly define expectations for its use, and what best practices and skills can make it an ally rather than a shortcut.
One particularly helpful idea is using GenAI to design class activities. While Kahoot keeps students engaged, their attention can still drift during lectures. Now, in each session of CIS 427, I dedicate 10–15 minutes to interactive activities that encourage students to think, discuss, and reflect on the topic. I asked GenAI to generate ideas for these interactions—some worked well, others not so much. I also had it generate reflection questions; again, the quality varied, so I now request many options and handpick those that fit best.
Interestingly, one new activity was inspired by GenAI, though not directly generated by it. I came up with the idea of a “network protocol charades” game, where students act out a protocol while others try to guess which one it is. It turned out to be a hit—students laughed, debated, and demonstrated their understanding in creative ways. This experience reminded me that while GenAI can be a great source of inspiration, human intuition and classroom experience are still essential for creating meaningful learning moments.
I also decided to bring back the IoT project. Previously, students had to figure out everything on their own in a domain they weren’t familiar with. Now, I teach them how to use GenAI effectively—to find answers, verify them, and apply them thoughtfully. This shift has worked well. GenAI allows students to explore more complex projects without depending too much on TAs or instructors, as long as they learn how to ask the right questions. Here are some videos from recent student projects: SmartDog, Intellitrash, and Smart Light. This time, the project was more open-ended: students could choose any sensors they wanted and build any system they wanted on Raspberry Pis, as long as it included a networked data transmission component.
Looking ahead, I often reflect on how what we teach must evolve in the age of GenAI. In the past, students had to memorize equations—how to calculate transmission delays, build routing tables step by step, or convert hexadecimal to decimal. But such tasks can now be easily handled by GenAI. What becomes more important, I believe, is a deep understanding of how networks actually work. This foundational knowledge enables students to collaborate more intelligently with GenAI systems—by asking the right questions and critically evaluating the answers. I truly believe we are at the dawn of a new era—the transition from the pre-GenAI age to the post-GenAI age. To embrace this change, as educators, we must reflect on what aspects of our teaching should evolve so our students can remain competitive and capable in the GenAI era. There is still a long way to go, but it is an exciting journey to be part of.
Dr. Zheng Song is an Assistant Professor of Computer and Information Science at the University of Michigan-Dearborn. He earned his second Ph.D. in Computer Science from Virginia Tech in 2020, where his research focused on computing and networks, data science, machine learning, optimization, and intelligent systems. He also holds a Ph.D. from Beijing University of Posts and Telecommunications (2015), where he specialized in wireless networking and mobile computing.
Featured image from Pixabay.
This post is a deliverable from the Hub’s 7-week GenAI Program
