Artificial Intelligence (AI) Guidelines

This document explains Kinship's stance on AI: what's considered acceptable use, what’s discouraged, and how to use tools like Claude and Midjourney effectively and responsibly. We understand that AI can be a complicated or divisive topic within creative spaces; our goal isn’t to take a moral stance on AI as a whole, but to define how it may fit in our community specifically. We view AI not as a substitute for human imagination, but as a collaborative assistant -- a tool that can make writing more accessible, especially for those navigating disability, neurodivergence, language barriers, or burnout.

Acceptable Use

You’re welcome to use AI for:

• Brainstorming ideas for characters or plots.
• Finding phrasing or structure when you’re stuck.
• Suggesting edits to, or proofreading, your writing for grammar, tone, or clarity.
• Generating prompts, summaries, or inspiration, especially when time or focus is limited.
• Creating visual references for your characters, items, or locations (e.g., through Midjourney).

When using AI this way, you should still edit, adapt, and refine the result so it reflects your voice, style, and intent.

Unacceptable Use

AI becomes a problem when it replaces personal effort or authenticity. Please avoid:

• Submitting full AI-generated posts or bios without revision.
  • AI tools often misunderstand tone, pacing, or emotional nuance, which can make collaboration feel disjointed. They also tend to “recap” or rephrase earlier dialogue and scenes instead of advancing them, resulting in repetitive or stalled threads. Each post should meaningfully move the thread along so both writers remain active participants, rather than one person providing all the momentum.
  • AI may also generate conflicting personality traits, add fantasy or supernatural details to appearances that should be earned or purchased, or create abilities and backgrounds that exceed what’s acceptable for the character’s race, level, or current progression. Always review and revise AI-generated content to ensure it fits within Kinship’s setting, systems, and standards before submitting.
• Using AI-generated text to impersonate another player’s writing style or character.
• Creating or using AI-generated images of another person's character without their permission, especially in any way that may make the person uncomfortable!

If you’re unsure whether something crosses the line, please just ask Support!

Using AI Writing Tools

Examples: Claude, DeepSeek, Gemini

• Use it as a partner, not a ghostwriter. Let it support your ideas, not replace your creativity.
• Be specific about tone and detail. If you just type “write me a bio/post,” it’ll produce generic filler. Include your goals, emotions, or setting cues so the result reflects your intent.
• Always review and rewrite. Don’t just accept the first output -- you should always edit pacing, dialogue, and detail until it feels authentic to your character and writing style.
• Fact-check carefully. AI may confidently invent details, misread context, or ignore prior events in a thread. Double-check everything it produces before posting.
• Ask it to help you learn. It can explain grammar choices, show formatting options, or suggest synonyms that better fit your tone or intent.
• Provide clear examples. Share snippets of your writing, mention what you like or dislike, or include stylistic references. The clearer your input, the more accurate the result.
• Stay within Kinship’s rules and systems. Use AI to enhance your creativity, not bypass established mechanics, lore, or progression limits.
• Protect your voice. AI can polish your writing, but don’t let it erase what makes your style unique. We want to write with you, not a machine.

Using AI Art Tools

Examples: Midjourney, Canva

• Use clear, descriptive prompts. Write what you want to see; instead of site-specific details AI won’t recognize, describe visuals directly, such as “slender red-haired man with gold eyes, wearing travel leathers, fantasy lighting.”
• Add style and composition cues. Guide the overall look by including terms like “digital painting,” “full-body portrait,” “standing side-by-side,” or “cinematic lighting.” This helps the AI match your intent for framing, pose, and mood.
• Avoid using specific artists’ names or signature styles. Prompts like “in the style of [artist]” can copy copyrighted work or artist aesthetics without permission. Instead, describe the qualities you want, such as “paint texture,” “soft lighting,” or “graphic novel illustration.”
• Refine through iteration. Expect to test multiple versions before you get a satisfying result. Adjust details like pose, lighting, and composition between generations -- guided trial and error often yields the strongest images.
• Expect post-generation cleanup. AI art frequently misrenders limbs, faces, or clothing and may produce incorrect colors or details. Plan to edit, crop, or touch up your results so they accurately represent your character or concept.
• Credit real artists when applicable. If your image includes non-AI elements, uses a base drawn or painted by another person, or incorporates art from an artist who’s asked for acknowledgment, always credit them clearly.

Final Thoughts

• AI can be a fantastic creative tool/supplement. It can enhance your process, speed up brainstorming, or help you write on low-spoon days, but your voice and intent are what make posts worth reading and responding to; it should never (and literally cannot) replace the human element in this process.
• If you have questions, concerns, or need assistance using AI, please don’t hesitate to reach out privately to Support! While we may not have all the answers, we’re happy to help where we can.
• If you suspect someone may be using AI inappropriately, please contact Support privately. We don’t support “AI witch hunts” or public accusations -- especially since many writing styles, including those of neurodivergent writers, are often unfairly labeled as AI-generated. Kinship values kindness and understanding above all else. Our goal is to guide members toward appropriate and creative uses of AI, not to punish honest mistakes or learning curves.
• Due to ChatGPT's involvement with the US Department of War, we can no longer recommend its use; please consider alternatives such as Claude or DeepSeek.