During the first week, I gave a lot of thought into the kind of project I wanted to create, and eventually settled on a main goal. I wanted to create an educational game, but not the kind that screams “this is a lesson”.

To me, a great educational game is one where players have fun, explore, and mess around... gaining knowledge without even realising it. If players leave with slightly more awareness than they started with, I will feel I have achieved something worthwhile.

The Challenge of Educational Games

Many educational games fail because they prioritise didactic content over genuine engagement. As Bogost (2007) argues in Persuasive Games, the most effective way to communicate ideas through games is not through explicit messaging, but through procedural rhetoric — the art of persuasion through game mechanics and rules themselves.

"Rusty machines don't always break down. Sometimes they keep chugging along, spewing out products of ever-diminishing quality."

This quote resonates deeply with my vision for Bird Corp. The “rusty machine” metaphor applies perfectly to how AI systems, despite appearing efficient, may be producing outcomes that gradually degrade environmental and social wellbeing. My game aims to embody this critique not through lectures, but through the player’s own choices and their consequences.

Rather than telling players that AI has environmental costs, I want the game mechanics themselves to reveal this truth organically through gameplay decisions and their rippling effects on the office environment and NPC relationships.

References

Bogost, I. (2007). Persuasive Games: The Expressive Power of Videogames. MIT Press.

Strubell, E., Ganesh, A. and McCallum, A. (2019). 'Energy and Policy Considerations for Deep Learning in NLP', Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pp. 3645-3650.

What Is Bird Corp?

Bird Corp is a short narrative serious game set in a fictional corporate office populated by anthropomorphic bird characters. The game starts out with the player stepping into their first day at work, meeting colleagues, attending a meeting, and running errands. However, throughout the game, the player picks up on things that don’t quite add up, and what began as a typical “first day at work” shifts into something investigative.

The core arc involves uncovering Bird Corp’s hidden secret. Near the end of the game, the player gains access to a restricted basement, which reveals the fact that the company has been drawing vast amounts of electricity and water to power the company’s AI systems. That moment is designed to reframe everything the player has been doing up to that point.

What the Game Is Trying to Do

The aim of Bird Corp is to raise awareness of the environmental costs of AI use in a way that feels experiential rather than instructional. Rather than reading about data centres or carbon footprints, the player lives inside a world where AI is completely normalised. The hope is that by experiencing that normalisation and then having it challenged, players come away with a different perspective on their own everyday AI habits.

The game is not trying to tell people to stop using AI. That would be unrealistic. Instead, it is about encouraging more deliberate and thoughtful use, and recognising that what feels invisible and effortless on the surface actually depends on a very real physical infrastructure.

The Academic Context

Bird Corp sits within the area of serious game design, which is games built with an educational or social purpose beyond entertainment. My research drew on frameworks in game design theory, player agency, and practice-based creative research to shape both the structure of the game and the process of making it. The game was developed iteratively over the course of the module, with each stage of building, testing, and reflecting feeding back into the next.

Thinking About Agency

One concern I had was whether a game with only one ending could still feel meaningful. I came across ideas around player agency that helped me work through this. The argument that stuck was that you don’t need total freedom for a game to feel purposeful. Agency can come through constrained involvement where the player feels genuinely entangled in the world rather than just moving through it. That reassured me that a guided narrative could still feel purposeful, as long as the player’s actions mattered within it. I also read about procedural rhetoric, the idea that the argument should come through what the player does rather than what they read. These became a core principle for how I built the mechanics.

Theoretical Framework: The Triadic Model

To ensure this project functions as a true Serious Game rather than dry “edutainment,”, the Harteveld’s Triadic Game Design Model ended up being the most useful framework I found that guided me throughout my game design (Harteveld and Kortmann, 2009).

This framework balances the core design elements, ensuring the experience remains engaging while delivering genuine educational value. The model consists of three essential elements that work together to create meaningful gameplay experiences: Reality, which connects players to real-world contexts; Meaning, which provides genuine learning opportunities; and Play, which ensures the experience remains immersive and enjoyable through engaging characters, storylines, and challenges.

Applying the Triadic Model to Bird Corp

In designing Bird Corp, I have deliberately mapped each element of the triadic model to specific game features:

Procedural Rhetoric: Learning Through Mechanics

Building upon the triadic model, I am also incorporating Bogost’s (2007) concept of procedural rhetoric — the idea that games persuade through their rules and systems rather than explicit messaging. As Bogost argues, games are uniquely positioned to make arguments through the processes they model.

In Bird Corp, the core "Manual vs. AI" choice mechanic embodies procedural rhetoric: by making AI the easier option with hidden long-term consequences, the game's rules themselves argue about the seductive danger of convenience over sustainability. Players don't read about this tension — they experience it through gameplay.

Triadic Game Design

The Three Essential Elements

References

Bogost, I. (2007). Persuasive Games: The Expressive Power of Videogames. MIT Press.

Harteveld, C. and Kortmann, R. (2009). 'Serious Gaming, an Introduction', in Triadic Game Design: Balancing Reality, Meaning and Play. Springer, pp. 3-19.

López, N. and Julio, P. (2020). 'Plot Twists as a Narrative Mechanic in Video Games', Game Studies, 20(1).

Risley, K. et al. (2025). 'Abstraction and Realism in Serious Games: Balancing Engagement and Learning', International Journal of Serious Games, 12(2), pp. 45-62.

Character Design

Once I knew the direction I was going for, I started with the Character Designs in Aseprite. Starting with a simple Budgie, I then felt the urge to put it in a suit, and from there started my spree of putting birds in different kinds of office wear.

inspiration for the birds

budgie first look vs final

more birds!

Environment Design

After I was done with the initial cast of characters and their basic animations, I began to work on the environment assets. I started with the tileset (floor, walls) and began to work from there

tileset_aseprite

plan for office env

toilet

The Initial plan

After making the basic assets, I began working on solidifying the game mechanics and storyline. I wanted this to be a game that was not just educational, but also fun and engaging. This involved designing relatable minigames in the form of office tasks, and meaningful Player-NPC interactions to keep players invested in the world of Bird Corp. I had to think of minigames that would be relatable in the office setting and able to be completed via either using AI or done manually.

Starting off in Godot: Characters and Maps

I went with Godot as the game engine mainly because it’s open source, has strong support for 2D games, and came highly recommended across game development communities. Since I was working solo without a game development background, knowing that I could turn to the plethora of online tutorials when I got stuck was really reassuring.

To get started with Godot, I searched for beginner tutorials on YouTube and followed along to build the basics of the game (player movement, tilemaps, and collisions). I used this tutorial for the parts I needed, specifically implementing the player scene, tilesets, and collisions, and learned that at the foundation of Godot games there are many node and scene types I could use.

player scene

player movement

beginning of map process

after decorating with furniture

After implementing the player movement, I had to implement the Tileset using TileMapLayers, creating a base for the map. Then returning to Aseprite to populate the map with decorations.

Cutscenes and Dialogue

After implementing the foundations (characters and environment), I had to start building the world of BirdCorp through narrative. While searching for how to implement cutscenes, I landed on Dialogue Manager by Nathan Hoad, but I could not understand how to implement it at first and continued searching. I came across the same video again later on, replaying it many times and found a section I missed. I was then able to set up Dialogue Manager properly.

A screenshot of my search history

After getting DialogueManager to work, I then wrote basic dialogue for each NPC according to my plan: Meeting the manager outside the toilet, then being led into the office for an introduction. I also had to implement functions in the NPC scene itself, for use through Dialogue Manager.

Manager.move_to_pos() # moves the npc to a specific x position
Manager.wait_idle() # sets the npc to idle state, facing a specified direction

Dialogue Manager script for the manager scene at the start

Meeting Notes Minigame

After creating Dialogue up to the point of where the minigame needs to be implemented, I started by designing the Minigame UI in Aseprite. This minigame was about taking meeting notes so it would be done on a laptop.

Minigame UI draft in Aseprite

Minigame Outcome

Screenshot of Meeting Notes Scene in Godot: Each sentence was thought out and separated into Labels in their own SentenceContainers.
The code then randomizes the Labels in each Sentence Container.

It was also at this stage where I started relying more on AI to aid me in coding. I used it to help implement the word rearranging and drag-and-drop feature. Players have rearrange randomized sentences into coherent sentences which will return visual and audio indicators when in the correct format.

Audio (SFX, BGM)

At some point, I also started implementing SFX such as the footsteps when walking, and character speech (learnt from DialogueManager tutorial). Anything that had meaning or action would be accompanied with a suitable sound.

I used Audacity to record for simple SFX like the footsteps, and character speech, and looked online for SFX such as door open/close and sink sounds. As for the BGM, I could not find a suitable one online so I generated it with Suno AI, the site that uses AI to generate music.

Supervisor Feedback

In one of my consultation session, Professor Conor McKeown suggested adding multiple endings and stronger narrative branching. The point was to give players clearer consequences for their choices, so the experience would feel more meaningful and replayable. I thought it was really valuable feedback, and I agreed with the reasoning. When different choices lead to different outcomes, the game feels more responsive to the player.

The main reason I did not implement that direction was time. Multiple narrative paths would have meant a lot more dialogue writing, state tracking, and testing, and I was already developing the project alone without a professional game development background. I had to be realistic about scope and focus on what I could complete to a quality I was confident in within the module timeframe.

After my first prototype, he also suggested implementing helpful visual indicators to guide players on what is interactable. Such as “E” to interact and a tutorial of sorts. I took in his advice and created a keybind tutorial showing all the main controls, and a floating E above interactables. Which I believe made a really big difference in the UX design.

keybind tutorial on top right and E to interact on an object

E to interact with NPCs

Updated Storyline and Mechanics

Instead of branching endings, I shifted towards building an illusion of choice. Players still make decisions throughout the game, choosing how to respond in conversations or whether to use BirdAI or complete a task manually.

I also explored different ways to improve the narrative:

“How to make finding the Basement natural”
“How to give characters more personality”
“How to add more lore to the world”
“How to strengthen the link to environmental impact”

For better flow, I implemented the idea that the player is a spy searching for a secret inside the company. That framing made the investigation feel motivated from the start, instead of feeling like a sudden plot turn.

For character personality, I gave each main character one defining word (sweet, anxious, lazy, etc) and then added extra animations to match those traits. This made their behavior feel clearer and helped interactions feel less flat.

For adding lore, I placed interactive items around the map that reveal short pieces of text-based lore, such as a newspaper, a crumpled paper, and a file.

For strengthening the link to environmental impact, I designed a power meter that rises whenever the player relies on BirdAI, with positive feedback and praise attached to that increase. At first, it was only meant to represent power usage as a gameplay system. Later, I realised it could carry a second meaning: the raise could be seen as praise and social levels, but could also later reveal to signal greater environmental impact due to AI usage. That dual meaning helped the mechanic support both player motivation and the core message of the game.

undercover introduction

personalized character animations

lore interactables

rising meter after AI usage

Added Content

Afterwards, I continued with the new concept and made more locations, characters, and dialogue. BirdCorp map now contains 3 levels and more other areas. Also adding collectables as well in order to progress past area locks to eventually reach the Basement.

map with a cut off section at the bottom

level 1

server room

janitor area

boss office

basement

My workflow during this process was mostly designing the maps in Aseprite, then forming them in Godot scenes. I would then prompt Copilot and tell it what I wanted to achieve such as implementing the Lift feature to travel between floors and forming the interactables.

My development process was like a continuous iterative loop: designing, programming, and troubleshooting via online resources and the use of AI. This repetitive cycle of playtesting to catch errors, and do fine-tuning became the rhythm of the project, allowing me to constantly refine the game.

Looking Back

Finishing Bird Corp and looking back at the whole process, I can see how far I have come. At the start, I was learning almost everything as I went. By the end, I had built a working and polished game, even if it is still a basic version of what the project could become. The hardest moments were when things did not work the way I intended, when AI prompts failed to give useful results, or when logical errors appeared and broke parts of the game flow. Those were the points where progress felt the slowest, but they were also where I learned the most. I am genuinely thankful for this experience because it pushed me to problem-solve, iterate, and trust the process, and I am equally grateful for the guidance and support from Professor McKeown and Supervisor Stuart Bell.

the development cycle behind Bird Corp

development challenges along the way

What Didn’t Fully Work

I think the honest limitation of the project is that it’s better at building some awareness than producing lasting behavioural change. Research into serious games suggests they’re good at shifting attitudes and increasing knowledge, but longer-term behaviour change is much harder to demonstrate. That matches my sense of what Bird Corp does.

What I Would Develop Further

There’s a lot I’d want to add with more time. More varied tasks, additional minigames with the manual versus AI choice built in, a deeper narrative, and a more polished set of animations and transitions would all make the game more compelling. Bird Corp feels very much like a prototype in the best sense of that word. It shows what the project can be, not the finished version of it. Further development would be a natural next stage, not a correction of failure.

Final Thought

The biggest thing this project taught me is that meaningful progress comes from persistence, not perfect execution. Bird Corp is not a flawless game, but it is a complete and thoughtful one, and it represents a huge step forward in my practice. I am leaving this project with stronger technical skills, clearer design instincts, and much more confidence in my ability to build ambitious ideas into something real.