Date
Publisher
arXiv
While game-based learning is widely used in programming education, few tools
offer adaptive, real-time support for complex topics, such as C pointers. We
present DeliverC, a GenAI-enhanced game that integrates GPT-4-mini to provide
personalized hints and generate pointer-related challenges on the fly. In a
pilot study involving 25 undergraduate students, we investigated the impact of
the system on learning through gameplay data and a 15-item survey that covered
constructs such as motivation, self-efficacy, metacognition, and feedback
quality. Results show that most students felt more confident and reflective
after using the tool, and error rates decreased as students progressed through
scaffolded levels. However, participation decreased with task difficulty, and
some students reported receiving unclear or vague feedback. These findings
suggest that DeliverC can enhance engagement and understanding in systems
programming, although refinement in AI-generated feedback is still needed. Our
study highlights the potential of combining GenAI with game-based learning to
support personalized and interactive practice in traditionally challenging
programming domains.
Study at a glance
- Question: How does an AI-powered learning game affect students' learning and motivation when studying computer programming?
- Setting and data: 25 undergraduate students
- Method: Pilot study analyzing gameplay data and a 15-item survey
Why it matters
Game-based learning is widely used in programming education, but few tools offer adaptive, real-time support for complex topics such as C pointers.
Key findings
- In a 25-student undergraduate programming course, most learners reported feeling more confident and reflective after using the AI game.
- Student error rates decreased as players progressed through the game's guided levels.
- Student participation declined as task difficulty increased, and some learners reported receiving vague or unclear AI feedback.
Summary drafted with AI from the paper’s abstract and may contain errors. See the full paper for details.
What is the application?
Who is the user?
Who age?
Study design
