GAI-A
GAI-A Research Project
Generative Artificial Intelligence for Architecture: AI-Assisted Co-Design Interface.
About the Project
GAI-A (Generative Artificial Intelligence for Architecture) is an AI-assisted co-design interface aimed at supporting creativity in architectural design processes. This interface enables designers to collaborate with an AI partner, encouraging creative thinking in the initial stages of the design process.
The integration of advanced information and communication technologies, especially in creative design fields, has revolutionized collaborative working processes by enabling stakeholders to work together from different locations and time zones.
GAI-A has been rigorously implemented and evaluated in the architectural design studio at Istanbul Technical University for four consecutive academic semesters.
GAI-A Interface Introduction Video
Our Team
Prof. Dr. Leman Figen Gül
Istanbul Technical University, Department of Architecture
Specialized in architectural design, computational design, and digital design, leading the GAI-A project.
fgul[@]itu.edu.tr
Burak Delikanlı
Istanbul Technical University, Department of Architecture and Graduate School
Continuing research on computational design and artificial intelligence applications.
burak.delikanli[@]itu.edu.tr
Oğulcan Üneşi
Istanbul Technical University, Graduate School
Playing an important role in the development of the GAI-A interface.
unesi23[@]itu.edu.tr
Ertuğrul Ömer Gül
University of New South Wales, School of Computer Science and Engineering
Contributing to the project with expertise in artificial intelligence and computer science.
e.gul[@]unsw.edu.au
Atacan Küçüksarı
Istanbul Technical University, Department of Architecture
Contributing as a member of the project research team.
kucuksari21[@]itu.edu.tr
Emin İşçi
Istanbul Technical University, Department of Architecture
Contributing as a member of the project research team.
isci21[@]itu.edu.tr
Publications
Exploring Co-design with an AI Partner: The GAI-A Interface in Architectural Education
Leman Figen Gül, Burak Delikanlı, Oğulcan Üneşi, Ertuğrul Ömer Gül
The integration of advanced information and communication technologies has revolutionized collaborative working processes, especially in creative design, allowing stakeholders to interact across locations and time zones. Originally designed to facilitate human collaboration, these systems have evolved to include collaborative interactions between humans and Artificial Intelligence (AI) models, reaching new heights with the integration of Generative Artificial Intelligence (GenAI) models. In this study, we introduce a co-design interface, Generative Artificial Intelligence for Architecture (referred to as GAI-A), for the initial design phase to stimulate creative design. GAI-A provides a collaborative space for designers to work with an AI partner that suggests visuals based on the images provided. Over the span of four consecutive academic semesters, we have rigorously implemented and evaluated this innovative environment in the architectural design studio at Istanbul Technical University, aiming to elucidate its multifaceted applications and implications within the academic discourse of architecture. The interface of GAI-A is evaluated through Creativity Support Index and surveys and semi-structured interviews. The research explores critical aspects, including co-design possibilities with GenAI models, and optimal utilization techniques, providing valuable insights into design pedagogy.
Can Artificial Intelligence Be a Creative Design Partner? Our Insights from the Architectural Design Studio Experience
Leman Figen Gül, Burak Delikanlı, Oğulcan Üneşi, Ertuğrul Ömer Gül
In recent years, there has been a growing interest in the use of deep generative models in the practice and education of architectural design. End-user applications of Generative Artificial Intelligence (GenAI) models are becoming widespread, and the impacts of these applications at the intersection of design and creative processes is a topic deserving of research. To explore these effects, for the past two years, we have been working with students at Istanbul Technical University, Faculty of Architecture, Department of Architecture in a series of design studios. In this sequential process, each semester we have built on what we learnt from the previous semester and collected information through questionnaires and focus meetings during and at the end of the studios, both to identify the status quo and to determine student expectations. As a result of our research and observations, we developed the GAI-A (Generative Artificial Intelligence for Architecture) platform, which is a user-friendly interface. According to our findings from interviews with students, critical aspects such as the integration of GenAI into design and optimal utilization techniques were examined, and its potential to foster creativity was investigated. This paper examines the application of the GAI-A platform in architectural design studios as a useful design interface, utilizing surveys and semi-structured interviews to evaluate its effectiveness, providing valuable insights into the current state of design pedagogy and AI-enhanced creativity.
GAI-A Interface
Interface Development Process
The integration of advanced information and communication technologies has revolutionized collaborative working processes, especially in creative design, allowing stakeholders to interact across locations and time zones. Originally designed to facilitate human collaboration, these systems have evolved to include collaborative interactions between humans and Artificial Intelligence (AI) models, reaching new heights with the integration of Generative Artificial Intelligence (GenAI) models. In this study, we introduce a co-design interface, Generative Artificial Intelligence for Architecture (referred to as GAI-A), for the initial design phase to stimulate creative design. GAI-A provides a collaborative space for designers to work with an AI partner that suggests visuals based on the images provided. Over the span of four consecutive academic semesters, we have rigorously implemented and evaluated this innovative environment in the architectural design studio at Istanbul Technical University, aiming to elucidate its multifaceted applications and implications within the academic discourse of architecture. The interface of GAI-A is evaluated through Creativity Support Index and surveys and semi-structured interviews. The research explores critical aspects, including co-design possibilities with GenAI models, and optimal utilization techniques, providing valuable insights into design pedagogy.
- Collage Tool: Users can place, rotate, resize, and arrange images on a canvas.
- Positive and Negative Prompts: Users can further condition generation with text.
- Advanced Options: Ability to adjust internal parameters of the sampler for advanced users.
- Output History: Users can access and compare their previous generations.
- Layer Ordering: Ability to arrange the order of images in the collage.
- Same-Seed Generation: Option to maintain seed values for consistent results.
GAI-A Collage Tool Interface
Interactive canvas for creating image collages with AI-powered generation
GAI-A Collage Tool Interface
Input Collage
Input Collage
Input Collage
Input Collage
User Evaluations
The GAI-A interface has been evaluated through the Creativity Support Index (CSI) and surveys. Research shows that 87% of students stated that GAI-A had a positive impact on their creativity in the design process.
Student Feedback
"
GAI-A was very helpful as a source of inspiration, especially in the early stages of design. It’s an excellent tool for visualizing abstract concepts and materializing complex ideas.
"
GAI-A was very helpful as a source of inspiration, especially in the early stages of design. It’s an excellent tool for visualizing abstract concepts and materializing complex ideas.
GAI-A by Leman Figen GÜL
- 0000-0001-9374-4620