Inteligência artificial generativa de imagens no ensino de arquitetura: impactos na criatividade e no processo de projeto

• Ano de defesa: 2026
• Autor(a) principal: Ourique, Gabriel Mustafa
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Santa Maria; Brasil; UFSM; Programa de Pós-Graduação em Arquitetura, Urbanismo e Paisagismo; Centro de Tecnologia
• Palavras-chave em Português: Arquitetura; Criatividade; Ensino; Processo de projeto; Inteligência Artificial Generativa; Architecture; Creativity; Generative Artificial Intelligence; Education; Design process; CNPQ::CIENCIAS SOCIAIS APLICADAS::ARQUITETURA E URBANISMO
• Link de acesso: http://repositorio.ufsm.br/handle/1/39191

This research investigates the use of Generative Artificial Intelligence (GenAI) imagegeneration tools during the preliminary design stage of the architectural design process, focusing on their effects on the creativity of Architecture and Urbanism students. Given the rapid dissemination of generative AI technologies, a mixedmethods (qualitative–quantitative) approach was adopted, structured around a methodology developed specifically for this study, entitled the Creativity Test with Reference Support (TCAR). The study was conducted in two experimental stages: the first involved the development of façade design proposals by second- and fourth-year students, who were divided into groups with different forms of access to visual references (GenAI, social media, websites/magazines, and a control group); the second consisted of a design activity using GenAI exclusively to develop inspiration boards and formal proposals for naturalized playgrounds. The assessments considered students’ self-evaluations and instructors’ evaluations based on the criteria of fluency, flexibility, originality, and elaboration (Torrance, 1966), as well as design functionality. The results indicate that GenAI enhances fluency, flexibility, and elaboration, but not originality. Students perceived the tool positively, whereas instructors provided more moderate evaluations. It is concluded that GenAI offers relevant contributions to architectural design education; however, critical mediation is required to prevent overreliance on the technology and encourage innovative solutions.