Research Articles at Koli Calling 2025
Summary of some of the many insightful articles about GenAI in CS Education at Koli Calling 2025
We’re still at the beginning (only three years!) of understanding how to integrate GenAI into CS courses. And while there is lots to do, we wanted to take this opportunity to highlight the amazing research that our community is doing.
At the recent KOLI 2025 conference, we observed many papers about incorporating GenAI into CS courses. But it’s the speed with which our community has advanced the research that is so impressive here. We’re going beyond should students use/not use GenAI, to focus on how students are using it, and how AI Can support them as they learn. We’ll just highlight three papers, but we encourage you to skim the proceedings for more.
Stephanie Yang, Melissa Chen, Bertrand Schneider
How do students use GenAI when seeking help? This study suggests that student motivation (“I want to understand”, “I want high grades”), and the value that students see in the tasks that we assign, has a lot to do with it!
Patrick Bassner, Anna Lottner, Stephan Krusche
This is a randomized experiment comparing Software Engineering students coding a compression-related algorithm with an educational AI tutor, ChatGPT, or no AI support. This study did not find significant differences in learning gains, completion time, or code accuracy. We need more research like this in courses beyond CS1!
Sonsoles López-Pernas, Kamila Misiejuk, Eduardo Oliveira, Mohammed Saqr
This is a great example of a new kind of analysis we’re seeing: studying the temporal/dynamic interactions of students with GenAI. This paper shows that students rarely engage in deep cognitive strategies (e.g. reflection, evaluation) when working with GenAI. We need to support students to become independent learners!
GenAI in CS Education Workshop: Practice and Research (AICSEPAR)
The inaugural GenAI in CS Education Workshop brings together leaders in research on the impact of GenAI on computing education and educators pioneering new curricula, pedagogies, and assessments through the integration of GenAI. Attendees should expect to learn about top research in the area as well as learn the practical details of GenAI integration. The research presented will be papers from top venues in the past year, with awards for top papers. Educators with experience integrating GenAI will apply to present.
Research Articles at Koli Calling 2025
Koli featured a number of insightful articles about GenAI in CS Education.
We’re still at the beginning (only three years!) of understanding how to integrate GenAI into CS courses. And while there is lots to do, we wanted to take this opportunity to highlight the amazing research that our community is doing.
Featured on UCSD News
Transforming Computer Science Education in the Age of AI
UC San Diego launches global consortium with support from Google.org to help teach the next generation of coders.
Educators, education researchers and computer scientists are teaming up as part of a consortium led by the University of California San Diego to reshape the future of computer science education in the era of generative AI...
CRA NAIRR AI Education Fellowship
Join CRA's Leadership in AI Education
The Computing Research Association (CRA), with support from the National Science Foundation (NSF), is pleased to announce the launch of the NAIRR AI Education Fellowship at CRA. The application portal is now live and will remain open through November 1, 2025.
Improving Student-AI Interaction Through Pedagogical Prompting
Prompt for Learning
Concerned about your students struggling to productively interact with
GenAI? Watching them one-shot solutions without learning much? Teach
them how to prompt for learning!
Pedagogical prompts are prompts that your students can use to
transform GenAI into an effective tutor. But many students likely do not
engage with GenAI in a pedagogically-meaningful way. Learn how to help them!
Andrej Karpathy: Software Is Changing (Again)
Software development hasn't changed much in 70 years. Now it has.
We've gone from Software 1.0 to 2.0 to 3.0, and students need to know how to operate at each of these levels. At the same time, we're just at the very beginning of GenAI-integrated coding: many of us are directly interacting through GenAI through prompts, not integrated GUIs! Watch Andrej Karpathy break down the shifts in software development due to GenAI.