top of page

AI workshop
AI, Game Design and Community

Problem

In 2025, I noticed that students were no longer strangers to AI compared to last year, when we first organized the AI camp. They now understand AI’s capabilities in text and image generation at a conceptual level. However, when it comes to their own creative processes, I rarely see them integrate AI tools. For instance, when searching for an image reference, students still prefer Pinterest over asking AI to generate more context-specific references.

​

Together with Safinah Ali and Ayat Abodayeh, I organized a three-day AI, Game Design, and Community workshop at the MIT Media Lab. Our goal was to study how students use AI during creative work. We chose game design as the medium because:

  1. It is inherently multi-media, involving visual art, sound, motion pictures, and coding—all of which can involve AI tools.

  2. It follows a complete creative cycle—initiation, pre-production, production, and testing—requiring students to collaborate and practice a range of skills.

  3. It's a favorable sandbox that students can imagine freely in terms of storytelling and worldbuilding in relation to real-world issues.

​

Our guiding questions were:

  1. How do students perceive AI-facilitated creativity during the game development process?

  2. How does AI enhance students’ capacity in art creation, and how do they evaluate its outputs?

  3. How does AI-facilitated game design empower middle school students to take computational action on community-based or social issues?

Approach

The workshop was a three-day, full-day (9am–4pm) summer program for nine middle school students and one high school student. Only one-third of participants had prior knowledge of Python.

Students were divided into five teams.

Game type: Narrative Game

AI tools introduced: Ren’Py (game engine), VS Code, GitHub Copilot, Suno, Runway, Gamma

Other tools: Canva, file-type converters

​

​​

Key curriculum phases

  • Day 1: Ideation & Story Development

    Introduction, community issue exploration, Ren’Py tutorial, storyboarding, scriptwriting, and flowchart creation.

  • Day 2: Development & Implementation

    Setting up projects, coding in Ren’Py, creating assets, and iterative testing.

  • Day 3: Refinement & Showcase

    Polishing games, creating game design documentation, and presenting final projects.

To support the process, we developed Arcade Game Maker as a tool for documenting game design.

Result

All five teams created a narrative game addressing an issue they cared about—ranging from adolescent mental health, human trafficking, and global warming to unequal distribution of social resources and waste classification.

  • All teams used Copilot in VS Code to generate and debug code.

  • One team chose not to use AI for art assets, instead creating characters, scenes, and animations in Scratch.

Reflection

  1. Intensity of production:The workshop was intensive for students to create a game production within a limited timeframe. On the second day, there was a student who texted me at 9:30 pm asking about a coding question. It demonstrates ownership over their projects and eagerness to solve problems.

  2. Storytelling autonomy: Students relied on their own voices to develop their stories. Even when their narratives lacked clarity, they trusted their own ideas—showing that for them, creativity came first. This suggests that for middle school students, imagination and personal expression remain central, with AI serving at most as a support rather than a substitute.

  3. AI in coding: Copilot empowered students to work at a creative level they could not have reached alone within such a short timeframe. Still, they lacked the technical knowledge to fully evaluate Copilot’s output, and often required teacher intervention to interpret or correct mistakes. This indicates that AI can accelerate learning and production but cannot yet replace the role of a human guide.

  4. AI image and video generation exposed practical barriers. The mismatch between AI-generated formats and game engine requirements forced students to rely on external tools (e.g., Canva, Photoshop, file converters). These extra steps slowed progress and created frustration. The integration—not just generation—is essential for AI to be meaningfully valuable for creative workflows.

  5. Selectivity in art: Character design was a key point of resistance: three teams chose to draw their own characters, while another expressed frustration at AI’s inability to meet their standards. This discernment likely reflects prior aesthetic training, suggesting that students do not passively accept AI outputs but actively evaluate them against personal expectations.

bottom of page