Week 9 – Project Reflections on the Now
This week has been a juggling act of the following courses: Text Remix, Gamifying Historical Narratives, Creative Visual Narratives, Interdisciplinary Futures, and my Final Project. As such, the aforementioned have overlapped to inspire each other – creating an amalgamation of sorts.
For example, the manhua comic I am drawing for CVN has bled into the art style I am utilising for character design in GHN. Both, combined, have and will inspire my final output for my Final Project as it is a visual piece. Furthermore, as my Final Project will be a visual board of a game design, the historical elements and game development in GHN is super relevant and inspirational at this point.
To answer the question – now what and why? – in regards to my project, I will analyse the following:
- How has your project developed?
Most importantly (at this point), I have identified the AI software I will be using: GPT-4. This makes the most sense as it is accessible through the UoE, and is readily accepted as ‘the most developed’ in its field. It also has the components necessary for my ‘experiment’, as detailed in the Scientific American article titled The Latest AI Chatbots Can Handle Text, Images and Sound. In the article, GPT is described as being able to effectively react to visual prompts provided. Specifically, “when given a photograph of a stocked bookshelf, [GPT] offered detailed descriptions of the hypothetical owner’s supposed character and interests that were almost like AI-generated horoscopes.” Thus, my focus in the following weeks will be in creating my ‘visual game design board’ to have the AI react to.
- What were some of the key ideas/texts/resources shared by your peers that moved your project forward?
My peers raised an interesting question as to whether it’d be best to have two different forms of AI react to the ‘visual game design board’, and then compare their responses. Others, in the same meeting, asked whether perhaps contrasting a human and AI response might yield more interesting results. After dwelling on both, I approached the course convener who advised either selecting two humans and one AI, or two AIs and one human. With all this in mind, I looked at research online and found that the most prevalent contemporary method seems to be a) pitch two AIs against one another or b) pitch an AI against a human participant. Therefore, I intend to conflate (and modify) both by having one source of AI and two human participants . The human participants will be an expert in game development and a newbie – this will allow for interesting results on which the AI’s results are ‘closer’ to.
- Where you’ve made changes, where did these decisions arise?
Originally, I was intending to use Miro to create my ‘visual game design board’ as it offers ‘infinite’ space and is a good visual software for presenting. However, after some preliminary testing with my GHN character designs, I realised that it wasn’t well-suited for a predominately ‘artistic’ approach and had limited options when it came to uploading additional components and editing them. Thus, I have chosen to work with Canva and Autodesk instead; drawing first of Autodesk and then editing the ‘visual board’ on Canva.
Therefore, the following week’s focus will be on: creating a historical game concept (different to the one I am currently carrying out for GHN). As of now, my team and I are looking at Revolutionary France meets Cyberpunk for our GHN project. As I already had a game concept prepared (think The Three Kingdom Period of Korea x The Three Musketeers), I will be looking into it with more depth and beginning to design an initial draft of my eventual ‘visual game design board’.