Would be nice if you could elaborate on how generative AI was used in your project.

Hi Criby,

Here's a detailed breakdown of the current generative AI elements within Handle:

Stable Audio 3 Stable Audio 3 was used to create the individual background tracks. I'm not great with music creation, so this allowed me to include the atmospheric music I'm going for. The background audio consists of 20 tracks spanning 90 minutes, with 3 variants of each track that use a random walk algorithm to create a non-repetitive score for the in-game day (60 tracks total). Additionally, it was used to generate 26 optional "bonus" tracks for the in-game MP3 player.  Additional post-processing was done for initialization and smooth track transitions.

Gemini & Python

• The 3 in-game wallpapers included with the J Desktop Environment, along with the logos, were generated by Gemini.
  • I fully expect players to upload their own wallpapers, and no default wallpaper is configured -- just a blue desktop background.
• Most of the in-game flavor text is procedurally generated. Some of the placeholder text was generated or edited using Gemini and the Python Faker package.

Sound Effects & Future Plans All sound effects currently in the game are code-generated (relating only to the telephone system) and do not use AI. However, I do intend to use AI generation for sound effects in the future, which is why I chose to include it in my AI disclosure. You can see this under the "More Information" section on the Handle landing page:

AI Disclosure AI Assisted, Graphics, Sounds, Text