I thought I won’t have much to reflect on in relation to my electives, but let’s try. Perhaps unfortunately, I have chosen the AI and Storytelling, hoping that I am gonna learn how to train AI to support me in juggling the multilayered stream of ideas and thoughts I have been dotting down in regards to my project, its narrative structure, game mechanics etc. I did not. But I will, on my own, somehow.

One “insight” that I guess could be gained from the AI elective was tied to how exactly that technology works (or at least worked in its early GPT-2 editions, and I assume the basics still apply now). Generating statistical predictions on what the next word/pixel is, based on available data, the technology is inherently bound to streamline “what’s out there”, in a feedback loop of eliminating everything that is not statistically mainstream and reinforcing that which is. The ethical aspects of this bias has obviously been explored to a greater depth also in the Ethical Data Futures, and somehow it finds its way also into every second elective there is, and into every second talk held outside of class over coffee or other beverages.

What I am carrying out of all this is a pervasive sense of pessimism and criticality towards AI, sometimes even echoing inside an imperative of AI avoidance or straight on boycott.

Yet, from the early designs of my project plan, I was always counting on that I would use AI somehow:

• to help me organise my scattered sets of ideas into a coherent game design;
• to simulate play tests for balancing, effectively taking on the role of other players for extra playtesting sessions (I cannot rely just on real people accepting my invitations for “playtest & chocolate” as many times as I’ll need it);

… and those uses seem, to me, even taking in the ethical considerations, quite alright, as long as AI still just presents a supplementary tool …

But then there’s the use of AI for a more generative purpose:

• to create game artwork for my prototype.

And here I am now a bit conflicted.

I focused my ambition for this project to create a game prototype, that is a game that can be played and further developed into a “final product”. What I see as key aspects of this project, is the way I translate ecosystem restoration into game mechanics, the way I use in-game narrative elements to achieve the effect on players that I desire.

And the art that’s on the surface of all this, I have thus far considered as “extra”, not so “academically important”, so to say.

I do not have the capacity, the talent, or the money/collaborators to make sure that my tree species card has a cool artwork on it that’s marvelous to look at (like, for example, the below artwork from Parks). Or game board tiles that have a really cute texture of forests. Etc etc.

So, I always counted that I will present a version of the game as my final project that will not have any of its art finalised. Which is alright, because the art is not really the most essential and will always be just a cosmetic to how the game works.

But I also don’t wanna present a bareboned game with blank cards and single-coloured boards, who’s gonna want to play that? I wouldn’t.

Sooo, is it alright that I generate images for all my trees and forests and related visuals, in a somewhat coherent art style? Who will I be exploiting, if I do so? But is there an ethical offset also? So many questions come to mind.

• Is there AI bias for trees and plant imagery the same way there is bias for people?
• Who am I potentially exploiting if I let AI generate an artsy art of a baobab, a monkey, or a forest?
• How do I capture my use of AI process in a way that is most transparent and “correct” towards those evaluating the project?
• Is it wrong to essentially “extract” from nature to develop a game that is all about redefining human-nature relationships towards a more integrated, ecosystem views?

But also, in a more positive aspect:

• Am I contributing positively to the various LLM’s machine learning loops if I generate large quantities of artwork specifically related to ecosystem restoration, endangered plant species, etc.? Essentially, asking it to do stuff that is not mainstream? After all, the mainstream view of ecology is still, to this day, primarily oriented at superficial large mammal icons like pandas, gorillas and rhinos – and I would really like nudge AI a little bit to imagine these topics with more diverse imageries than just that of a “stereotypical” Amazon rainforest with a burning orangutan.
• Am I enabling myself to focus more on the stuff that matters most to me in this project – mechanics and narrative – by outsourcing the prototype artworks to a generative AI?

 

I’ll leave it at those questions for now, and see how they ferment in the next couple of days or weeks. But I’m eager to hear anyone’s thoughts on the matter. Since the artwork will be in any case one of the last stages of the project development, I do not need to worry about this for a long while. Although, putting some provisional artwork on it already during development will certainly make the whole creative process a bit more pleasing.

 

Feb 13, 2025