Where do games disclose AI use?

Some background music tracks were generated with the assistance of generative AI during development. All AI-generated music has been manually reviewed, edited, and verified by the developer to ensure it does not infringe on any copyrights or intellectual property. The application itself does not include any generative AI functionality.

The static background map of the campaign was generated by an AI. All texts were localized by AI – all languages ​​except English and German.

We utilised AI for concept art and UI assets.

AI tools have been used to generate part of the localization texts.

We used LLMs to help consolidate the rules engine code into one module for offline and online play. We also worked with Epidemic Sound, whose voice actors use AI to create voice content with their own voices.

AI is used to generate some images for the store page and in-game.

Background pictures

Placeholder AI-generated unit voice-lines in current playtest build, no AI usage targeted for final release build.

Some card faces are generated using AI

Generative AI has been used to generate graphics for the game, these include Buildings, Wonders and Resources.

I use commercially licensed AI generated graphics assets as temporary placeholders for some card images and icons because of the frequent card design changes. They will be replaced by manual art after the card designs are finalized in the released game.

The character concept art and skill icons are AI-assisted creations. AI has helped us produce a considerable amount of 2D assets.

I used local art generation (tiles, background) and audio generation (music).

Marketing materials. Title screen.

AI is used in gametesting and upscaling images for Steam store

A few images/icons were generated by AI

I have used AI to translate the game to French and have then tried to clean it up afterwards.

The game utilizes a mix of traditional audio and AI-generated sound effects.

Some character portraits and visual elements were created with the assistance of generative AI tools as part of the art production process. All assets were curated, edited, and integrated to maintain a consistent artistic direction. As the game continues development, certain visual elements may be refined or replaced. AI-assisted tools may also be used to support translations.

Our team leverages generative AI to elevate the creation of our graphical assets. This cutting-edge technology streamlines our workflow, making the process more efficient, enjoyable, and innovative. By utilizing generative AI, we can explore new artistic directions and deliver visually stunning experiences to our players.

The backgrounds and card icons in this game are created using AI.

Some assets on the Steam page were created using AI

Some royalty-free music used in the game was sourced online and may involve the use of AI.

AI has been used for creating various icons for the game, as much as possible is created by hand.

Beneath The Six features an AI narrator who serves as a dungeon master to guide you and comment on your experiences in the daily generated content. We commissioned actor Tom Canton (The Witcher TV series) to record a library of dialogue. This was then used by ElevenLabs to train the AI voice. Both Magicave and ElevenLabs have a commitment to ensuring the rights of voice actors are protected. Magicave has assigned ownership of the AI voice to Tom Canton, licensing it from him for use in the game, while allowing him the freedom to use his AI voice in other media – whether games, voiceovers, animations, audiobooks or movies (https://www.gamewatcher.com/news/beneath-the-six-turn-based-roguelike). Each day we generate new daily content for the dungeon and arena. As part of this we pass the generation data to an LLM to create dialogue for the AI narrator so they can comment on the content included for that day. Some of the daily generated content for the arena and dungeon modes use an LLM to create names for two other purposes. One is for creating the name of any Champion enemies which appear in the dungeon. The other is for creating a name for the daily arena run.

Much of Fire Isle uses Adobe Firefly generated art as a base for the UI buttons. I've further edited/photoshopped all of this art personally to better suit the game's art style and make it my own. As a broke solo dev without much of a 2D art background and no funding to hire anyone, I chose to go this route rather than to not see Fire Isle made at all. Despite the reasoning behind it, I understand many might not be comfortable with AI generated content. I hope the game itself reflects the thousands upon thousands of hours of time & effort I've put into it's creation with the goal of making a fun and memorable experience, NOT AI Slop🙏...

- Voice overs are created using artificial intelligence.

Using AI tools for language translation

At the time we launched in Early Access, item art, UI art, and 2D environment art was AI-sourced. We have now replaced all of these with hand-drawn art. Please note that, as of writing, the trailer on the store page shows a previous version of the game with AI-generated UI assets, though it is still highly representative of what the game currently looks like. Please view the screenshots at the top of the page and the gifs on the store page to see what the game looks like currently with no AI-generated assets in it.

The promotional and waiting images of the game use generative AI to make the work more complete. The actual content of the game lacks generative AI

We used AI to fill in details and at times generate images for some of the store pages. Additionally, the music that is in the background while playing the game was generated using AI tools and services. In all such cases, the final product reflects the creativity and expression of our own development team.

This game contains assets that may have been generated by AI. The developer cannot precisely identify which assets are AI-generated.

AI-generated voiceovers.

Initial localization for non-english languages was done using large language models. The community has improved on that localization immensely. The game itself contains no generative content, with the exception of localization.

Part of arts are AI-generated content and part of scripting are supervised by ChatGpt.

images

Marketing content - logos, posters, banners are generated using DALL-E. Some of 2d images ui, figures and 3d models are generated using AI. Also few sounds are generated.

-Images created with artificial intelligence

LLM was used to translate the text into different languages.

Game uses pre-rendered AI generated images for 2D Art.

Generative AI was used to partially create the background images on the marketing assets.. The game itself is hand-crafted, hand-drawn, and animated by a human hand!

This game uses AI assistance to create some of the 2D art assets.

As a solo dev, I use Midjourney to fill in some gaps including upgrade and achievement icons.

As a solo developer, I use AI to brainstorm drafts for some of the 2D assets, which I then edit by hand. These will be replaced with hand-drawn assets by release.

Critias Empire uses images of public domain 18th and 19th century classical oil on canvas paintings, but since there are only a limited number of these, AI generated images are used where no appropriate historical public domain image can be found. This includes backgrounds, illustrations for buildings, features, disasters, maps and characters, plus live ops, events and marketing material. The images are generated using paid subscription to Midjourney. Crystalline Green Ltd. owns the copyright on the generated images. We are a small independent studio with limited budget, but we appreciate not everyone is comfortable with use of AI generated images in commercial games, so we understand if you would prefer not to play our game. Feel free to contact us if you have further questions or concerns.

AI generated images have been used to as a starting point for icons, characters, and elements of some backgrounds. In nearly every case images are then digitally painted over to remove unwanted artifacts and give all the artwork a similar style. Every effort has been made to produce unique content that does not infringe on any copyright.

Some dialogue backgrounds and icons use AI generation.

Some icons and background images are currently prototyped with AI but will be replaced in the future if the budget allows it.

This game used AI to help quickly iterate through designs for some UI elements using our own artwork as a baseline. AI is also used to help translate the game; and is used for the AlienGPT custom species system using player inputs.

Usage of AI-Generated and Assisted Illustrations in Shadow Empire Post-release Illustration Creation Process The developers use Midjourney and Nano Banana to generate high-resolution illustrations for new functionality added post-release, particularly for Vidcom and Stratagem artwork. These images are often post-processed by hand, as well as enhanced using tools like Magnific, Topaz Labs, or other upscaling and refinement software. AI in Post-release Portrait Art Portraits: A base image is generated using Midjourney with a generic prompt such as “portrait of young woman, military.” Smaller parts (eyes, nose, mouth, hair, etc.) are overlaid using DALL·E to create variations. Aging effects are achieved by rendering an older skin texture with partial transparency over the younger base face. The older the character, the more opaque the overlay becomes. (Note: Eyes and lips often retain a youthful appearance.) This effect is gradually applied using a limited number of sprite layers, allowing nuanced visual transitions. AI-Generated Art in Post-release Game Content The game includes AI-generated and AI-assisted illustrations for Vidcoms and Stratagems. All such AI-generated art is pre-generated — no generative AI tools are called or used in real-time by the game itself. The intention is to provide players with an option to toggle between the original and updated AI-enhanced graphics via preferences. (Note: This is an aim and may not yet apply to all content.)