Datastel: Moontlike nut van ChatGPT 4-o in juweliers ontwerp klaskamers, faktore wat die gebruik daarvan kan dryf en riglyne vir varantwoordelike gebruik.
收藏Mendeley Data2026-04-18 收录
下载链接:
https://data.mendeley.com/datasets/rxxtn25wh3
下载链接
链接失效反馈官方服务:
资源简介:
This dataset documents an exploratory, iterative series of ChatGPT-4o interactions (prompts P4–P108) and the associated AI-generated jewellery designs, descriptions, and researcher interpretations that form the empirical basis of an article. It primarily comprises table-organised outputs on: (1) ChatGPT-4o’s “baseline” designs without guidelines and the influence of material specification on stylistic preferences; (2) the effect of stylistic guidelines (e.g., Art Nouveau/Art Deco) and the inclusion of techniques on originality, stereotyping, and potential cultural/stylistic appropriation; (3) theme-based prompts (social impact, environmental issues, violence against women and children) showing how prompt refinement changes narrative, focus, and design quality; (4) experiments using “foreign/non-existent contexts” (extraterrestrial narratives) to stimulate original aesthetics; (5) ChatGPT-4o’s ability to visualise and “improve” rough sketches (with the risk of normalising designs into stereotypical solutions); and (6) AI-generated manufacturing plans (digital and traditional) that are logical but often lack sufficient technical detail. Collectively, the dataset indicates that ChatGPT-4o is particularly useful for ideation and conceptual development, provided that an expert user guides it through specific, unconventional, and iterative prompts to manage feasibility, originality, and ethical risks.
本数据集记录了一系列探索性、迭代式的ChatGPT-4o交互(提示词(prompt)编号P4–P108),以及与之配套的AI生成珠宝设计、设计说明与研究者解读内容,构成了某篇学术论文的实证基础。本数据集主要包含以表格形式整理的以下六类输出:(1)无指导规则下的ChatGPT-4o“基准”设计,以及材料规格对风格偏好的影响;(2)风格指导规则(如新艺术运动(Art Nouveau)、装饰艺术(Art Deco)风格)与工艺纳入对原创性、刻板印象,以及潜在文化与风格挪用的影响;(3)基于主题的提示词实验(涵盖社会影响、环境问题、针对妇女与儿童的暴力等主题),用于展示提示词优化如何改变叙事方向、研究焦点与设计质量;(4)使用“外来/不存在语境”(外星叙事)激发原创美学的实验;(5)ChatGPT-4o的可视化与“优化”手绘草图的能力(同时存在将设计同质化至刻板方案的风险);(6)AI生成的制造方案(涵盖数字化与传统工艺),此类方案逻辑自洽但通常缺乏足够的技术细节。综合来看,本数据集表明,若由专业用户通过针对性、非传统且迭代式的提示词引导ChatGPT-4o,以管控可行性、原创性与伦理风险,则该模型在创意构思与概念发展阶段尤为实用。
创建时间:
2026-02-04



