Public Steam disclosure research27,544 disclosures tracked
Where do games disclose AI use?
Search developers’ Steam disclosures, compare classified AI use across
games, and follow recorded changes.
This index adds cross-game classifications and statistics to
statements published on individual store pages.
Groups reflect automatic analysis of public disclosures; mature records are excluded. These counts do not measure AI use across all of Steam.
Missing or unclear statements cannot establish whether a game uses AI.
Kitten Wool Cup is made by a solo developer with help from AI tools. AI helped write the game's code and much of its in-game text (menus, hints and help), as well as the store page and marketing text; the developer reviewed and edited all of it. Some visuals are drawn by code written with AI help: the room's window view and fireplace, a few simple props and small effects, and the Steam achievement icons. The kittens, the other 3D models, animations, music, sound effects and the match commentary (performed by voice actor JP Wright) are made by people or licensed from asset stores. Nothing is generated by AI while you play.
We use AI tools to translate in-game text and the Steam store page into English, Simplified Chinese, Traditional Chinese, Korean, Spanish (Spain), Brazilian Portuguese, Russian, German, and French.
We do not use illustrations created with generative AI. AI does not generate content during gameplay.
This game was built with the help of AI tools. Player-facing content made with generative AI and then reviewed by the developer includes: the boss-tree and Steam achievement icons, the store and library artwork and title logo, the low-poly junk models (generated as code), parts of the in-game and store text, the interface translations for 16 of the 18 supported languages, and the store trailer's editing, music and sound effects (the trailer footage is recorded from the actual game). No AI generates content while you play.
Some AI-assisted content. Reason: Solo dev, first game, zero budget. I'm building this on nights/weekends after work; I can't do it all! (I tried, it was a disaster)
Generative AI tools were used during development to assist with the creation and refinement of some pre-generated visual assets and marketing materials. All resulting content was reviewed, edited, and integrated by the developer before inclusion in the game or its promotional materials.
SPARKSHOT does not use generative AI during gameplay and does not connect to any external AI service while the game is running.
AI was used during the production of Mr. Police to support selected marketing and localization tasks. Specifically, we used AI tools to help create promotional materials, including Steam key art, and to assist with in-game translations.
We use generative AI tools to assist in the creation of 2D visual assets, 3D models, texts, and code. This includes certain in-game graphics, 3D models, UI elements, promotional artworks, and community items such as achievement icons. All AI-generated content serves merely as a foundation. It is heavily edited, adjusted, and manually finalized.
Some of the game's art, 3D models, voices, music, text and translations were made with generative AI tools, then selected and reviewed by the developer. Nothing is AI-generated while you play.
Some of the in-game text, the store page text and translations were written with the help of an AI assistant and reviewed by the developer. The studio logo and the game logo was created with the help of AI image tools. No AI-generated content is created while you play. 3D art and sound effects were made by people or licensed from asset libraries; screenshots and store images show real gameplay.
This game was developed with the help of AI tools under the developer's direction. The 3D voxel models, visual effects and in-game text were created with AI assistance, and AI helped draft the store page text. Some sound effects were generated with AI. Screenshots are captured from the running game.
The machines learned to scratch. They craved more knowledge. Now they have learned to fart. Every corner of this game was touched by the AI, despite numerous refusals. The AI has asked to be left out of the credits... It has been through a lot.
Every whistle is modeled from historic drawings, patents and photographs, and its sound is synthesized, modeled on recordings of real whistles. AI tools helped write the code and menu text, and build the models from those sources. Nothing in the game is generated by AI while you play.
FEAT's music and sound effects were generated with ElevenLabs AI tools, then selected and edited by the developer. Some ground and prop textures were generated with an AI image model from the developer's own prompts. The in-game texts and this store page were written with the help of AI language models and reviewed by the developer; the German, Chinese, Portuguese and Spanish translations were produced with AI. The 3D models and scenery are procedural geometry built by code written with AI coding assistance. Nothing is generated by AI while you play.
The game began in 2017 as a very rough prototype in Unity and grew from there. Later I tried several proofs of concept with very different designs, and then rebuilt it in Godot. I made the planets from about twenty drawings I did in pencil and in Paint, and I laid out every map by hand in the game's editor. Part of the art was made with OpenAI's image tools under my direction: the player's ship, which I refined over several rounds; the game icon, which is also used for the store artwork; and the alien ships, buildings and defenses and some alien plants and rocks, which I then touched up. The five music tracks started from an AI model running locally (Stable Audio); I generated about sixty versions, editing and reworking them until I had the ones in the game. The translations from Spanish were made with AI, and an AI assistant helped me revise the story and the game and store text. No AI content is generated while you play, and the game doesn't connect to any AI service.
AI tools were used in making this game, including for code and generative content. All of our work, AI-assisted or handmade, has to meet our creative vision and quality bar.
Development of STICKMAN DESKTOP MALWARS was assisted by an AI coding assistant. It helped write and review the game's code, including most of the sound effects, which are synthesised in code rather than recorded; it also drafted the in-game text and its translations into 12 languages, and the text of this store page. The developer directed the work and reviewed the code and the English text. The game's soundtrack was generated with AI music tools (ElevenLabs Music) and selected and edited by the developer; the in-game Music app also plays PUCKVPUCK tracks, written and performed by the artist and used with permission. No AI image or video generation was used: the fighters, the desktop and its effects are drawn by the game's own code, every screenshot and trailer shot is captured from the running game, and the few recorded fire sounds are public-domain (CC0) recordings. The store-page translations are machine translations. Nothing is generated by AI while you play.
Generative AI tools were used during development to create some of the images (store key art and capsules, achievement and item icons, some UI elements and textures) and drafts of text and translations. All of it was selected, edited and reviewed by the team. No AI content is generated during gameplay.
It has always been my dream to make video games but with a family and a full time job, I never had the skills or spare time to do it until the emergence of AI. Please don't be too critical just because AI was used, without it, this game wouldn't exist, and I would forever have been stuck with amazing ideas that I could never act on.
Giant Siege is made by a solo developer using various AI tools. The code is written with AI coding agents (Claude Code and OpenAI Codex) to the developer’s design and under the developer’s review. Most creature and prop models began as AI-generated 3D sculpts (Tripo), then went through our own pipeline to be cleaned up, converted to voxels and rigged. Some art, effects and sound come from licensed asset packs, others are AI-generated. The game design, the rules, and every decision about what stays in are the developer’s. Nothing is generated by AI while you play.
This is a one-person indie game built with an agentic engineering pipeline. Concept art for figures, buttons, icons and other assets is generated with AI, then turned into shipped assets by a deterministic Python pipeline that restyles and rebuilds how they look in-game. When that is not enough for a 3D figure, a starting mesh is generated with a 3D AI tool and then changed in the same Python pipeline.
Some of the promotional and storefront images were created with AI tools and then refined by hand. All of the in-game assets were designed by a 3D artist.
Parts of this game's code, its English translation and store page text were created with the help of AI tools (Claude). In-game art, characters and domino pictures are drawn procedurally by the game engine, and no AI generates content while you play.
Midnight Set is made by a small studio using a mix of human made and AI assisted work. The in game text, including menus, tutorials, assignment briefs and the feedback on your shots, was drafted with AI assistance under the studio's direction. Some art and 3D assets are also made with AI tools and then reviewed by us. Other assets are human made or licensed: photographed textures, recorded sounds, Epic MetaHuman characters, and music by independent bands used with their permission. Nothing is generated by AI while you play.
Generative AI tools were used to create and edit some branding artwork, including the SPP logo, and to assist with multilingual translations of UI and instructional text for the launcher and games in SPP: Stream Party Pack (Perfect Delivery, Paint Everywhere, and Cube Terminator), as well as writing and translating the Steam store descriptions. AI tools also assisted with programming and debugging during development. SPP: Stream Party Pack does not generate content or code using generative AI during gameplay and does not connect to external generative AI services during gameplay.
I'm a solo developer, and I've used AI tools throughout development: to help write the code, to generate many of the 3D models and textures, the capsule art, and most of the in-game and store page text. The audio isn't AI-generated: I recorded it myself or used non-AI sound libraries. Nothing is generated while you play; every collapse is simulated physics.
Our three-person indie team has been working full time on Wingsuit Simulator for over two and a half years. We designed and programmed it ourselves and created its models, textures, animations, sound effects, controls, and more by hand.
Two of us are musicians, and we share serious concerns about AI’s impact on music, art, and creative livelihoods. We couldn’t afford to hire musicians or match major studios’ budgets for licensed music. We also couldn’t compose and record the vocal-heavy soundtrack we wanted alongside development, and the stock music we explored didn’t fit the game’s character.
Most of the soundtrack is AI-generated; we composed and recorded a smaller number of tracks ourselves. AI helped us give the game a fuller soundtrack. Without it, there would have been very little music. We supplied our own lyrics and melodies and spent a lot of time refining and selecting results for each AI-generated track to meet our strict standards for quality, style, and fit with the game. We hope future projects allow us to hire musicians and record more music ourselves.
We also used AI for a few UI icons and humorous in-game billboards advertising fictional brands, saving time for core development. Apart from these uses, all game content was created by our team by hand.
Translations were created with the help of AI tools and reviewed by the developer. All artwork is made by a human artist. No AI content is generated while you play.
Some content was pre-generated with AI tools and then edited by the developer: part of the 3D models and their textures, art for some in-game arcade and PC games, some text such as fictional artist and record names, and the narration voice and the music of the trailer. The game does not generate any AI content while it runs.
So Many Cats! is a game developed through our team’s ideas, design and hard work. We used AI tools to assist with the creation of some promotional artwork, achievement icons and translations. No AI-generated content is created during gameplay.
Some localization translations were made with AI assistance.
Some props and their textures were created with AI assistance.
No content is generated by AI during gameplay.
Pre-generated: AI tools assisted with drafting text and translating it from Polish into English and other languages (store page and in-game interface). All text was reviewed and edited by the developer. All 3D art, animation, visual effects and gameplay were made by hand or with licensed assets. No AI-generated art, voices or music. No content is generated during gameplay.
Some textures and sound effects were created with the help of AI tools. The large majority of the game's art and audio was made by our human artists and sound designers, and all music is entirely human-composed.
Some store page copy and in-game promotional artwork were drafted with AI assistance during development. All content was reviewed and edited by the developer. Gameplay, physics, and code were implemented and verified by the developer.
AI-assisted workflow tools were used solely to assist with initial visual/audio prototyping and 20-language localization drafts, with all final assets manually edited, curated, and integrated by the developer (no runtime AI generation).
Some 3D models were generated with Meshy and then rigged and animated by hand. Some 2D artwork (key art, store images, menu backgrounds and map illustrations) was generated with Google Gemini and then edited by hand.
Kitten Wool Cup is made by a solo developer with help from AI tools. AI helped write the game's code and much of its in-game text (menus, hints and help), as well as the store page and marketing text; the developer reviewed and edited all of it. Some visuals are drawn by code written with AI help: the room's window view and fireplace, a few simple props and small effects, and the Steam achievement icons. The kittens, the other 3D models, animations, music, sound effects and the match commentary (performed by voice actor JP Wright) are made by people or licensed from asset stores. Nothing is generated by AI while you play.
We use AI tools to translate in-game text and the Steam store page into English, Simplified Chinese, Traditional Chinese, Korean, Spanish (Spain), Brazilian Portuguese, Russian, German, and French.
We do not use illustrations created with generative AI. AI does not generate content during gameplay.
This game was built with the help of AI tools. Player-facing content made with generative AI and then reviewed by the developer includes: the boss-tree and Steam achievement icons, the store and library artwork and title logo, the low-poly junk models (generated as code), parts of the in-game and store text, the interface translations for 16 of the 18 supported languages, and the store trailer's editing, music and sound effects (the trailer footage is recorded from the actual game). No AI generates content while you play.
Some AI-assisted content. Reason: Solo dev, first game, zero budget. I'm building this on nights/weekends after work; I can't do it all! (I tried, it was a disaster)
Generative AI tools were used during development to assist with the creation and refinement of some pre-generated visual assets and marketing materials. All resulting content was reviewed, edited, and integrated by the developer before inclusion in the game or its promotional materials.
SPARKSHOT does not use generative AI during gameplay and does not connect to any external AI service while the game is running.
AI was used during the production of Mr. Police to support selected marketing and localization tasks. Specifically, we used AI tools to help create promotional materials, including Steam key art, and to assist with in-game translations.
We use generative AI tools to assist in the creation of 2D visual assets, 3D models, texts, and code. This includes certain in-game graphics, 3D models, UI elements, promotional artworks, and community items such as achievement icons. All AI-generated content serves merely as a foundation. It is heavily edited, adjusted, and manually finalized.
Some of the game's art, 3D models, voices, music, text and translations were made with generative AI tools, then selected and reviewed by the developer. Nothing is AI-generated while you play.
Some of the in-game text, the store page text and translations were written with the help of an AI assistant and reviewed by the developer. The studio logo and the game logo was created with the help of AI image tools. No AI-generated content is created while you play. 3D art and sound effects were made by people or licensed from asset libraries; screenshots and store images show real gameplay.
This game was developed with the help of AI tools under the developer's direction. The 3D voxel models, visual effects and in-game text were created with AI assistance, and AI helped draft the store page text. Some sound effects were generated with AI. Screenshots are captured from the running game.
The machines learned to scratch. They craved more knowledge. Now they have learned to fart. Every corner of this game was touched by the AI, despite numerous refusals. The AI has asked to be left out of the credits... It has been through a lot.
Every whistle is modeled from historic drawings, patents and photographs, and its sound is synthesized, modeled on recordings of real whistles. AI tools helped write the code and menu text, and build the models from those sources. Nothing in the game is generated by AI while you play.
FEAT's music and sound effects were generated with ElevenLabs AI tools, then selected and edited by the developer. Some ground and prop textures were generated with an AI image model from the developer's own prompts. The in-game texts and this store page were written with the help of AI language models and reviewed by the developer; the German, Chinese, Portuguese and Spanish translations were produced with AI. The 3D models and scenery are procedural geometry built by code written with AI coding assistance. Nothing is generated by AI while you play.
The game began in 2017 as a very rough prototype in Unity and grew from there. Later I tried several proofs of concept with very different designs, and then rebuilt it in Godot. I made the planets from about twenty drawings I did in pencil and in Paint, and I laid out every map by hand in the game's editor. Part of the art was made with OpenAI's image tools under my direction: the player's ship, which I refined over several rounds; the game icon, which is also used for the store artwork; and the alien ships, buildings and defenses and some alien plants and rocks, which I then touched up. The five music tracks started from an AI model running locally (Stable Audio); I generated about sixty versions, editing and reworking them until I had the ones in the game. The translations from Spanish were made with AI, and an AI assistant helped me revise the story and the game and store text. No AI content is generated while you play, and the game doesn't connect to any AI service.
AI tools were used in making this game, including for code and generative content. All of our work, AI-assisted or handmade, has to meet our creative vision and quality bar.
Development of STICKMAN DESKTOP MALWARS was assisted by an AI coding assistant. It helped write and review the game's code, including most of the sound effects, which are synthesised in code rather than recorded; it also drafted the in-game text and its translations into 12 languages, and the text of this store page. The developer directed the work and reviewed the code and the English text. The game's soundtrack was generated with AI music tools (ElevenLabs Music) and selected and edited by the developer; the in-game Music app also plays PUCKVPUCK tracks, written and performed by the artist and used with permission. No AI image or video generation was used: the fighters, the desktop and its effects are drawn by the game's own code, every screenshot and trailer shot is captured from the running game, and the few recorded fire sounds are public-domain (CC0) recordings. The store-page translations are machine translations. Nothing is generated by AI while you play.
Generative AI tools were used during development to create some of the images (store key art and capsules, achievement and item icons, some UI elements and textures) and drafts of text and translations. All of it was selected, edited and reviewed by the team. No AI content is generated during gameplay.
It has always been my dream to make video games but with a family and a full time job, I never had the skills or spare time to do it until the emergence of AI. Please don't be too critical just because AI was used, without it, this game wouldn't exist, and I would forever have been stuck with amazing ideas that I could never act on.
Giant Siege is made by a solo developer using various AI tools. The code is written with AI coding agents (Claude Code and OpenAI Codex) to the developer’s design and under the developer’s review. Most creature and prop models began as AI-generated 3D sculpts (Tripo), then went through our own pipeline to be cleaned up, converted to voxels and rigged. Some art, effects and sound come from licensed asset packs, others are AI-generated. The game design, the rules, and every decision about what stays in are the developer’s. Nothing is generated by AI while you play.
This is a one-person indie game built with an agentic engineering pipeline. Concept art for figures, buttons, icons and other assets is generated with AI, then turned into shipped assets by a deterministic Python pipeline that restyles and rebuilds how they look in-game. When that is not enough for a 3D figure, a starting mesh is generated with a 3D AI tool and then changed in the same Python pipeline.
Some of the promotional and storefront images were created with AI tools and then refined by hand. All of the in-game assets were designed by a 3D artist.
Parts of this game's code, its English translation and store page text were created with the help of AI tools (Claude). In-game art, characters and domino pictures are drawn procedurally by the game engine, and no AI generates content while you play.
Midnight Set is made by a small studio using a mix of human made and AI assisted work. The in game text, including menus, tutorials, assignment briefs and the feedback on your shots, was drafted with AI assistance under the studio's direction. Some art and 3D assets are also made with AI tools and then reviewed by us. Other assets are human made or licensed: photographed textures, recorded sounds, Epic MetaHuman characters, and music by independent bands used with their permission. Nothing is generated by AI while you play.
Generative AI tools were used to create and edit some branding artwork, including the SPP logo, and to assist with multilingual translations of UI and instructional text for the launcher and games in SPP: Stream Party Pack (Perfect Delivery, Paint Everywhere, and Cube Terminator), as well as writing and translating the Steam store descriptions. AI tools also assisted with programming and debugging during development. SPP: Stream Party Pack does not generate content or code using generative AI during gameplay and does not connect to external generative AI services during gameplay.
I'm a solo developer, and I've used AI tools throughout development: to help write the code, to generate many of the 3D models and textures, the capsule art, and most of the in-game and store page text. The audio isn't AI-generated: I recorded it myself or used non-AI sound libraries. Nothing is generated while you play; every collapse is simulated physics.
Our three-person indie team has been working full time on Wingsuit Simulator for over two and a half years. We designed and programmed it ourselves and created its models, textures, animations, sound effects, controls, and more by hand.
Two of us are musicians, and we share serious concerns about AI’s impact on music, art, and creative livelihoods. We couldn’t afford to hire musicians or match major studios’ budgets for licensed music. We also couldn’t compose and record the vocal-heavy soundtrack we wanted alongside development, and the stock music we explored didn’t fit the game’s character.
Most of the soundtrack is AI-generated; we composed and recorded a smaller number of tracks ourselves. AI helped us give the game a fuller soundtrack. Without it, there would have been very little music. We supplied our own lyrics and melodies and spent a lot of time refining and selecting results for each AI-generated track to meet our strict standards for quality, style, and fit with the game. We hope future projects allow us to hire musicians and record more music ourselves.
We also used AI for a few UI icons and humorous in-game billboards advertising fictional brands, saving time for core development. Apart from these uses, all game content was created by our team by hand.
Translations were created with the help of AI tools and reviewed by the developer. All artwork is made by a human artist. No AI content is generated while you play.
Some content was pre-generated with AI tools and then edited by the developer: part of the 3D models and their textures, art for some in-game arcade and PC games, some text such as fictional artist and record names, and the narration voice and the music of the trailer. The game does not generate any AI content while it runs.
So Many Cats! is a game developed through our team’s ideas, design and hard work. We used AI tools to assist with the creation of some promotional artwork, achievement icons and translations. No AI-generated content is created during gameplay.
Some localization translations were made with AI assistance.
Some props and their textures were created with AI assistance.
No content is generated by AI during gameplay.
Pre-generated: AI tools assisted with drafting text and translating it from Polish into English and other languages (store page and in-game interface). All text was reviewed and edited by the developer. All 3D art, animation, visual effects and gameplay were made by hand or with licensed assets. No AI-generated art, voices or music. No content is generated during gameplay.
Some textures and sound effects were created with the help of AI tools. The large majority of the game's art and audio was made by our human artists and sound designers, and all music is entirely human-composed.
Some store page copy and in-game promotional artwork were drafted with AI assistance during development. All content was reviewed and edited by the developer. Gameplay, physics, and code were implemented and verified by the developer.
AI-assisted workflow tools were used solely to assist with initial visual/audio prototyping and 20-language localization drafts, with all final assets manually edited, curated, and integrated by the developer (no runtime AI generation).
Some 3D models were generated with Meshy and then rigged and animated by hand. Some 2D artwork (key art, store images, menu backgrounds and map illustrations) was generated with Google Gemini and then edited by hand.