Perception of AI and Human Collaboration

Research Hypothesis
The study hypothesizes that design students' perceptions of authorship in AI-assisted design vary depending on the degree of influence their original image prompts have on the final AI-generated outcomes. It aims to understand at what point students consider AI-generated images their own creation and how comfortable they are in collaborating with AI tools, giving credit to the AI where due.

Data Collection &Analysis
Data was gathered through a mixed-methods approach involving an AI-aided design workshop followed by a survey. Participants included second-year industrial design (ID) and interior architecture (IA) students. The workshop introduced the Vizcom AI tool, which students used to generate AI-assisted designs based on their studio project outcomes. The survey assessed students' familiarity with AI tools, perceptions of authorship, reasons for their perceptions, and comfort levels in co-designing with AI. The sample size was 30 students (12 ID, 18 IA). difference in familiarity level with AI tools (Q1), perception of authorship (Q2), reasons of perception of the puthorship (Q3), and comfort level of collaboration with ai (Q4) between ID and IA students. A Spearman's rank correlation coefficient can be calculated to assess the relationship between students' familiarity with AI tools (Q1) and their comfort levels in co-designing with AI (Q4). Textual analysis can be conducted to thematically analyze the responses from the open-ended question, Q5.

Notable Findings

T-Test: The T-test results indicate that there is no statistically significant difference between ID and IA students in terms of their familiarity with AI tools, perception of authorship, reasons for their perception of authorship, and comfort level with collaborating with AI.

Perception of Authorship: 33.3% of students considered AI-generated images as their own creation up to a 50% image prompt influence level. Significant thresholds for authorship were identified between 40% and 70% image prompt influence, with most students feeling a sense of ownership within this range.

Comfort with AI Collaboration: 46.4% of participants were comfortable or very comfortable collaborating with AI and might give credit to AI. A moderate positive correlation (rs=0.48) was found between familiarity with AI tools and comfort in co-designing with AI.

Reasons for Perceived Authorship: The primary reason cited was that the design was still based on the students' original research or findings (60%).

Implications for Future Research
The study's findings suggest a need for further research with a larger and more diverse sample. Exploring perceptions of more experienced designers and different design fields could provide a broader understanding of authorship in AI-assisted design. Future studies could also delve deeper into the reasons behind neutral attitudes towards AI collaboration and explore the impact of storytelling in enhancing AI-human collaboration.

研究假设
本研究提出如下假设：设计专业学生对人工智能（AI）辅助设计中的作者身份认知，会因其原创图像提示词对最终AI生成成果的影响程度而异。本研究旨在明确学生在何种阈值下会将AI生成图像视为自身创作，并了解他们与AI工具协作的接受度，以及如何恰当地为AI标注其贡献。

数据采集与分析
本研究采用混合研究方法，先开展AI辅助设计工作坊，随后进行问卷调查。参与者为二年级工业设计（Industrial Design，ID）与室内建筑学（Interior Architecture，IA）学生。工作坊介绍了Vizcom AI工具，学生依托自身课程项目成果，使用该工具生成AI辅助设计作品。问卷调查评估了学生对AI工具的熟悉程度、作者身份认知、该认知的成因，以及与AI协同设计的接受度。本次研究的样本量为30名学生（12名工业设计专业，18名室内建筑学专业）。
本研究将对比工业设计与室内建筑学学生在AI工具熟悉度（Q1）、作者身份认知（Q2）、作者身份认知成因（Q3）以及与AI协作接受度（Q4）上的差异。可计算斯皮尔曼秩相关系数（Spearman's rank correlation coefficient），以评估学生对AI工具的熟悉程度（Q1）与AI协同设计接受度（Q4）之间的关联。针对开放式问题Q5的作答，可开展主题式文本分析。

核心发现
独立样本t检验结果显示，工业设计与室内建筑学学生在AI工具熟悉度、作者身份认知、作者身份认知成因及与AI协作接受度方面，均无统计学意义上的显著差异。
作者身份认知：33.3%的学生认为，当图像提示词影响程度达50%时，可将AI生成图像视为自身创作。研究明确了作者身份认知的关键阈值区间为40%至70%的提示词影响程度，多数学生在此区间内会产生创作归属感。
AI协作接受度：46.4%的参与者表示愿意或非常愿意与AI协作，并会为AI标注贡献。研究发现，学生对AI工具的熟悉程度与AI协同设计接受度之间存在中等程度的正相关（rs=0.48）。
作者身份认知成因：学生提及的首要原因为，设计创作仍基于自身的原创研究或成果（占比60%）。

未来研究启示
本研究结果表明，未来需采用更大规模、更多元化的样本开展进一步研究。探索资深设计师及不同设计领域的作者身份认知，可深化对AI辅助设计中作者身份问题的理解。未来研究还可深入剖析学生对AI协作持中立态度的成因，并探究叙事性表达对提升人机协作体验的影响。