Using customized GPT to develop prompting proficiency in architectural AI-generated images

2025-04-16

AI Summary

This study addresses architecture students’ weak prompt engineering skills in generative AI–driven image creation. We propose a pedagogical intervention integrating interactive, anthropomorphized AI agents with a structured prompt guide. Grounded in reverse-engineering tasks—where students reconstruct prompts from given images—and leveraging a fine-tuned GPT model within a mixed-methods design, the intervention systematically develops students’ conceptual articulation, critical thinking, and prompt precision. Our key contributions include (1) the first application of personified AI roles in architectural education and (2) a novel multidimensional evaluation framework assessing prompt similarity, specificity, and lexical count. Experimental results demonstrate statistically significant improvements (p < 0.05) for the intervention group across all three quantitative metrics; additionally, qualitative analysis reveals marked gains in students’ prompt confidence and conceptual clarity.

Technology Category

• Natural Language Processing: Prompt Engineering / Prompting
• Humans and AI: Game Design — Procedural Content Generation & Storytelling
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

This research investigates the use of customized GPT models to enhance prompting proficiency among architecture students when generating AI-driven images. Prompt engineering is increasingly essential in architectural education due to the widespread adoption of generative AI tools. This study utilized a mixed-methods experimental design involving architecture students divided into three distinct groups: a control group receiving no structured support, a second group provided with structured prompting guides, and a third group supported by both structured guides and interactive AI personas. Students engaged in reverse engineering tasks, first guessing provided image prompts and then generating their own prompts, aiming to boost critical thinking and prompting skills. Variables examined included time spent prompting, word count, prompt similarity, and concreteness. Quantitative analysis involved correlation assessments between these variables and a one-way ANOVA to evaluate differences across groups. While several correlations showed meaningful relationships, not all were statistically significant. ANOVA results indicated statistically significant improvements in word count, similarity, and concreteness, especially in the group supported by AI personas and structured prompting guides. Qualitative feedback complemented these findings, revealing enhanced confidence and critical thinking skills in students. These results suggest tailored GPT interactions substantially improve students' ability to communicate architectural concepts clearly and effectively.

Problem

• Enhancing prompting skills in AI-generated architectural images
• Improving critical thinking via structured GPT-guided prompting
• Evaluating AI persona impact on prompt quality metrics

Innovation

• Customized GPT models enhance prompting proficiency
• Mixed-methods design tests structured guides and AI personas
• AI personas boost prompt similarity and concreteness significantly