AppsinityAPPSINITY
Policies and safety

AI generated content

AI-assisted work is allowed as long as you hold the rights to what it produced and you declare it. The rating form has explicit labels for AI chatbots, image generation, voice generation and AI generated content, so an undeclared AI feature is a misdeclaration rather than an oversight.

Open Appsinity Studio

Declare the AI you ship

FeatureRating label to tickWeightAlso declare
In-game chat assistantAI Chatbot2Data shared with the model provider
Player-facing image generationAI Image Generation2User Generated Content if outputs are shared
Synthesised voicesAI Voice Generation2Language and content of transcripts
AI-made art, text or music in the buildAI Generated Content1Rights to the training output
AI moderation of player contentAI features plus moderation0Moderation tools in Final Questions
Model that stores promptsAny of the above-Personal info, Device info as applicable
Watch out: The weights are small because AI presence is a disclosure, not a risk category. The disclosure is nonetheless mandatory: an AI feature is a claim about what the software does.

Rights are the hard part

The content rules ban stolen assets and unlicensed copyrighted material, and they do not offer an AI exemption. If a generated image resembles a protected character, if a voice model imitates a real person, or if a tool's terms forbid commercial use and you ship commercially, the violation is yours.

  • Keep the prompt and the tool name for anything you publish; that is the only realistic provenance record.
  • Read the tool's licence for commercial use before it becomes your business model.
  • Do not use AI voices that sound like an identifiable person.
  • Do not train on a specific artist's style and market the resemblance.
Not available: There is no AI provenance field, no watermark verification and no model registry on Appsinity. Your declaration is the record, and your licence is your own paperwork.

Machine text left in listings

The most common AI-related problem on the store is not the artwork, it is unedited generated listing copy. Two things give it away, and both hurt you: text that references another game, and text that still contains the scaffolding from the generation tool.

Real examples of scaffolding that has appeared in published pages: :contentReference[oaicite:3]{index=3}, [insert screenshot here], Certainly! Here is the description:, and markdown asterisks that render as literal characters.

  1. Generate a draft if you like. It is a fine first pass for structure.
  2. Rewrite every factual claim. Every feature you mention must exist in the build you upload.
  3. Delete the scaffolding. Paste the output into a plain text view and read it as a player would.
  4. Check the character limits. The short description is 80 characters; a generated sentence rarely fits without editing, see Short description.

AI on the platform side

Appsinity itself uses Gallifai, an AI-assisted safety system, to scan uploaded and submitted builds for malware, harmful software, suspicious files and deceptive builds, and to answer first-line support questions within a fixed topic scope. Its scanning reduces risk without guaranteeing detection, and it does not read your source or your metrics.

Note: See Malware policy for what the scanner actually inspects and Contact support for where a human follows up.

Frequently asked questions

Can I publish a game whose art is entirely AI generated?

Yes, if you have the rights to use the outputs commercially and you tick AI Generated Content in the rating form. Style imitation of a protected character remains a copyright problem.

Must I label AI content to players?

Declare it in the rating form and in the description. There is no automatic AI badge on a listing.

Do you reject listings with obviously generated text?

Not automatically. Unedited machine text is a review risk and, where it misrepresents the build, a metadata violation.

Are AI chat features restricted?

A chatbot inside a game is fine, and the same content rules apply to what it says. A chat feature that lets strangers talk with no reporting mechanism is the risk, see Content policy.

Need help?

Developer publishing support: [email protected]
General support: [email protected]
Partnerships: [email protected]