Has anyone tried AI for game development share your experience

germa59

New Member
Nov 3, 2021
7
4
13
Has anyone tried AI for game development share your experience

what have you used this for
1) asset creation
2) environment creation
3) Animation creation
4)programming
5) any thin else

Share your experience
 

osanaiko

Engaged Member
Modder
Jul 4, 2017
3,437
6,604
707
1) yes - takes a lot of repeated editing to get consistent characters, so while it's far better than my own near-zero drawing ability, it still takes what feels like forever
2) yes - used it quite a bit, and works well for my needs if based on 3d renders or real world photos
3) yes - just played with this a little, but the tools are getting better very quickly.
4) no - I am a programmer already, don't need it to write code for me.
5) yes - used local text-generation-webui with the storybook plugin for idea refinement and some writing tasks early on.

My tl;dr: it's helpful in some aspects but you MUST put in a tonne of your own efforts to get quality results, otherwise the output looks just like all the other ai slop. I only use the tools because I am a hobbyist and don't want to put money in to the project. If I was doing game commercially, i would use real artists as the results would be faster and better.
 

n00bi

Active Member
Nov 24, 2022
670
734
217
I am not a game dev per say, but i have played around with misc ai tools, not ones that generates visual stuff tho,
But LLM's (ollama), TTS and STT (text to speech and speech to text) etc

1) Yes - As in voice and music generation
2) Nop
3) Yes - Although not in the way you think.
4) Yes - But it still needs a human overview of the code it generates and you need the skills to understand coding.

5) Visme project.
I made a tool that utilize different AI tools. "F5", for voice generation, "WhisperX" to extract words, along some custom code to
extract and get phonemes and to create a visme track for characters (genesis 8.x).
It also uses ollama to analyze the text to give facial expression. sad, happy so forth.
You don't have permission to view the spoiler content. Log in or register now.
 

Mispeled

Newbie
Game Developer
Oct 30, 2020
43
60
37
Yes on all count.
1) Absolutely, it's what it's best at honestly. 2D mostly for me because I have only basic expertise in digital art.
2) Yes, in two ways: For generating 2D backgrounds and for placing other assets in 3D scene, especially when it comes to tedious stuff like decorating houses or creating forests and the like, which it does better than typical procedural generation techniques.
3) Yes, I use it in two ways: For motion capture and for general advices when I'm manually doing animations, either in 2D or 3D
4) Oh I'm a senior software engineer and I use AI for coding. Not using it is just bad in my opinion. Like using notepad over literally any IDE. It's a powerful tool. It can cut down on workload tremendously by generating structs, classes, and general code monkey stuff you'd hire an intern to write for you because you can't be fucked to waste precious time manually doing everything. I also use it for more complex stuff by having it generate code for stuff and then reading through the generated code to fix the issues I find. It's basically like having a free intern that never sleep. You code review the AI code and get what you want most of the time, and when you don't you just do it yourself anyways.
5) Yes. I use it for cooking recipes, general inquiries, training routine, etc.
 

red-snapper

New Member
Jun 4, 2025
2
1
3
I was full AI hater untill I started using it for art like 2 weeks ago.
Well I would hate it even more now if I was an illustrator.

1) I can make 512 images like a small factory and my card does not even have 8GB.
I think some of the results are great.

The real problem seems to be the hate some people will have for your game.
I can't even post info on my tentacle game in the biggest tentacle community as if
their main unifying principle was hatred for AI and not love for tentacles. My plan
was to gain enough traction to be able to hire real artist but well thant won't happen.
 

jai_investigator

Dev of Opening to Love
Game Developer
Apr 30, 2023
212
225
72
I am making a game with AI image right now.

1) yes - all my character images and H-scenes are AI generated. I've found good prompting and models help with consistency and quality. There's a learning curve to use it effectively, and to get good results, but I find once you get the formula right you can generate a lot of good quality images quickly.
2) kinda - If you mean backgrounds I use it a bit but I've found just getting free assets off itch etc. to be a lot better than what I can generate.
3) yes - It took a lot of time to find a good way to get consistent, quality results. I tried many different methods and it helped me learn a lot. I have now introduced short looping animations to the latest version of my game and the feedback has been positive so far.
4) no, not really - Maybe I might use Chatgpt to figure out how I might do a certain complex function but I'm pretty confident with basic Ren'py coding so it's really more like a google search and a helpful way for me to learn some things.
5) Nope not really.
 

MissFortune

I Was Once, Possibly, Maybe, Perhaps… A Harem King
Respected User
Game Developer
Aug 17, 2019
6,910
10,689
872
I am not a game dev per say, but i have played around with misc ai tools, not ones that generates visual stuff tho,
But LLM's (ollama), TTS and STT (text to speech and speech to text) etc
What do you think is the best current TTS for realistic voices. I've been using e2f5/xtts with ElevenLabs, but wondering if there's anything better I'm not aware of.
 

n00bi

Active Member
Nov 24, 2022
670
734
217
What do you think is the best current TTS for realistic voices
Oh that is hard to say as they all are getting good these days.
I do use F5 TTS and it works ok for the most part, and its free + runs locally.
However you need to provide it with voice samples it can "emulate".. it needs some references.
You can make dialog with misc charters. or use the multi speech function to create misc variation of same voice.
ie happy, sad, contempt etc. but you need a reference voice for all the voice "moods".
Some pre work should also be done to get it good.. normalize the sound volume for all voices etc.

Link

Example.
Arnold schwarzenegger
Kate Mulgrew (Captain janeway from star trek voyager)
Patrick Stewart (Captain Picard from star trek)
Katey Sagal (Leela from futurama)
Jonathan Frakes ( Commander William Riker from star trek )