AI-Asset Pipeline Ethics / Legal Workflow

Draft our studio's generative-AI asset pipeline policy: 1) Allowed uses: ideation/concept exploration, rapid thumbnails, texture/detail reference bases, internal audio prototyping, marketing mockups; disallowed: final shipping player-facing art without significant human transformation/ review, direct use of named living artists' styles in marketing, generated voices of real people, copyrighted character likenesses. 2) Provenance: log tool, model version, prompts, source images/license per asset in a registry; retain exports; final assets re-created or substantially transformed by named artists who sign off. 3) Legal hygiene: review each tool's commercial terms monthly (they change), no uploads of confidential art to public models, employee training, indemnification notes, platform store disclosure requirements for AI-assisted content (Steam questionnaire etc.). 4) Quality: art-director gate, consistency bible checks, manual touch-up pass, copyright-similarity checks. 5) Team/union relations: frame as tooling that increases concept throughput, not replacement; produce the final one-page policy and asset-intake form. ================================================================================ END OF VOLUME 3 · 22 PROMPTS 游戏制作类英文提示词合计 3 个文件 · 62 条 ================================================================================