About AI Usage in my projects

In the time of AI there are games that still are -->mostly<-- made by apes...

So, what does that "mostly" mean, exactly?

The usage of AI in Exoroids

I think it is important to be transparent about the use of AI in my projects, so let me explain how I use AI in my projects and what I think about AI in general.

Tl;dr / Summary

Game- Design / story / the vision and everything Meta is 100% me
3D Models are handmade by myself in Blender
Backdrops, Menues and most in- Game Graphics are handmade with Inkscape and Gimp
For specific purposes like the stars and nebulae in the background I use specific assets from the Asset Store to solve specific problems that would be to time consuming to create myself
Sounds are self recorded or come by asset packs, basically how it is done everywhere. Tools: FMOD, Reaper, Audacity
Music are mainly placeholders from the asset store
AI is used for:

Speech samples for the enemies AND the cutscenes
The portraits of the story characters.
Basic code assistance

Further plans regarding AI Usage:

I'll keep / overhaul the speech samples
I will try to find someone human to create the portraits of the story characters, if I know what exactly is needed in the end
I will try to find someone that makes music, but will also use AI to generate music to see if that is a viable option

Long version

I started this project in 2021, that was far before the advent of large language models like Chat GPT in the end of 2022. And even then, the usage of AI in programming or game development was new and untested, and most people adapted quiet slowly. So, the base of "everything Exoroids" is 100% handmade, including the low poly 3D Models I do myself in Blender.

The sounds you hear are from large Sound packs that were largely modified or combined with tools like Reaper, Audacity and FMOD. Some of the sounds are self-recorded, but to be honest, there is not that much need for custom sounds in this game :-)

The music is a complete placeholder that I plan to replace with new content, as are the Portraits of the Characters, because I suck at drawing, at least for the graphics I have in mind.

Of course, I use LLM for code assistance, those tools are too powerful to ignore, but most of the code I'm writing myself and architectural design decisions are made by me. No vibe coding, this is not planned for this project.

So regarding the things that I possibly WOULD do in the future with AI:

Music
Speech

-> Why music and speech?

The speech samples (especially the in-game voice reactions of the enemies) are a "nice to have". The quality of the generated speech must not be perfect since it is heavily distorted with custom effects to imitate space communications, so I think I just keep the voice-overs and redo the cutscene samples because I'm not satisfied with the quality.

The music is another beast...

I don't know if I find someone that will make custom, good-fitting music for very little money, as my budget is nearly zero, but I will try. If not, then I have literally not that much choice: Using assets from the asset store or buy finished music packs and hope to find something that fits OR using AI generation. It's too early to say what I will do, but I will definitely try AI-generated music out to make myself a better picture of those technologies and the quality they produce.

Some general thoughts about AI in games

Gaming is about emotions, and emotions are very personal and subjective. When we are playing games, we are not just playing the games, we often connect to the people and stories behind the games in some way.

Have you ever read the book "Masters of Doom"? It's "the" book about the development of "Doom" and the people behind it. It's a great read, and I highly recommend it, but that's not the point here.
The point is that there are stories behind the production of those games, insane stories sometimes about people doing people things and people connecting to or disconnecting from each other. There are stories about how certain features were developed or found by happy little accidents. Or stories how people destroy doors with swords.

What stories will you tell when you vibe coded a game?

Of course, that doesn't mean that a generated game may not be good. In the very near future the big companies AND most of the indies certainly will use the LLM generation MASSIVELY, because otherwise they won't be able to compete anymore, it's as simple as that. And the controversy if you SHOULD generate "Art" for your project will quickly fade. So it will be a question of conscience only if you use AI or not.

I bet we will see a lot of generated AI Slop soon and that it will be hard to find the good stuff within it. But I guess we will see some sort of counter-movement that will promote only 100% handmade games. So it won't probably be that bad. And the rest, like me, will be somewhere in between.

The debate over whether it is morally justifiable to use a technology based on the exploitation of a huge quantity of creators is, in terms of its impact, essentially already obsolete. The genie is out of the bottle, and there’s no putting it back in. As a society, we must ask ourselves in the future how we want to approach art, what human labor is worth to us and how we intend to deal with “data inflation.”

But: Of course, everyone is free to decide for themselves that using AI-generated content is out of the question for them and to boycott that 'crap'.

Let's see how it goes, and sorry for the long text ;-)