Learning by generating games with LLMs, supported by prompting pedagogy¶

Strategy · All strategies
Evidence · 4 claims (1 for, 1 mixed, 2 against) · 1 study (1 theoretical), q1 · 0 of 1 report an effect size · 4 claims rest on one study

Description¶

The article proposes that students learn educational content by designing and making games with generative AI tools, extending constructionist gaming. The technicalities of programming can be largely outsourced to LLMs so students focus on topic and game design. The authors coin "prompting pedagogy" for the pedagogical practices involved, including preparing prompting templates and sequencing prompting activities into phases such as idea generation, core design, prototyping, and assessment, integrated into a meaningful teaching process such as the creative thinking spiral.

Design Implications¶

Context¶

Requirements¶

• Students need support to develop adequate prompting skills; teachers should generate at least one learning game with an LLM before applying the approach.

Constraints¶

• Design and production activities need to be integrated into a meaningful teaching process; critical evaluation of generated outputs requires domain expertise.

Target Learners¶

• students in game-making activities
• teachers orchestrating LLM-based game-making projects

Target Learning Goals¶

• academic content learning
• prompting and critical evaluation skills
• digital fluency

Claims¶

• Well-designed games can promote persistence in practice and high-quality engagement, and when basic needs of autonomy, competence, and relatedness are met, learning becomes its own reward, reducing the temptation to outsource cognitive effort to LLMs. [+W]
• The convenience and low opportunity cost of LLMs poses a risk of over-reliance, deskilling, and reduced judgment competence when learners outsource tasks that are essential learning opportunities. [-W]
• Educational game adoption is hindered along socio-cultural, educational, and technological dimensions, including social rejection, limited accessibility to games and technology, and high development costs. [-W]
• Optimizing prompts can vastly improve the accuracy of LLM outputs, but critical evaluation of the resulting outputs still requires domain expertise. [~W]

Related Strategies¶

• Educators gamify learning materials or create learning games using generative AI with iterative, phased prompting
• When integrating GenAI, deliberately weigh control-oriented structures against designs that promote learner agency and creativity

Examples¶

-

Key Sources¶

• Huber, S. E., Kiili, K., Nebel, S., Ryan, R. M., Sailer, M., & Ninaus, M. (2024). Leveraging the Potential of Large Language Models in Education Through Playful and Game-Based Learning. Educational Psychology Review. https://doi.org/10.1007/s10648-024-09868-z