This thesis analyzes the role of the prompt in the use of generative artificial intelligence for image creation, with a particular focus on text-to-image models. After an overview of the origins of AI, the work explores the limitations of these systems, especially algorithmic bias, visual hallucinations, and issues related to copyright. A central part of the research is devoted to prompt engineering, understood as a strategic methodology for guiding generative output through accurate and effective textual design. The thesis aims to identify techniques and approaches that enhance the quality of generated results by fostering a creative interaction between user and machine. The final section presents a practical case study: the design of an illustrated calendar developed in collaboration with the real estate agency “Tetto Rosso”, using a workflow that integrated tools such as ChatGPT, Midjourney, and Photoshop. This project demonstrates how the conscious and critical use of language in prompts can become a powerful expressive tool for designing and producing original visual content.
La presente tesi analizza il ruolo del prompt nell’utilizzo dell’intelligenza artificiale generativa per la creazione di immagini, con particolare attenzione ai modelli text-to-image. Dopo una panoramica sulle origini dell’IA, si approfondiscono i limiti di questi sistemi, in particolare i bias algoritmici, le allucinazioni visive e le problematiche legate al diritto d’autore. Una parte centrale del lavoro è dedicata al prompt engineering, inteso come strategia per guidare l’output generativo attraverso una scrittura progettuale precisa ed efficace. La tesi si propone di individuare tecniche e approcci capaci di aumentare la qualità dei risultati generati, valorizzando l’interazione creativa tra utente e macchina. Il percorso si conclude con un caso studio applicativo: la realizzazione di un calendario illustrato in collaborazione con l’agenzia immobiliare “Tetto Rosso”, sviluppato mediante un flusso di lavoro che ha integrato strumenti come ChatGPT, Midjourney e Photoshop. Questo esempio dimostra come l’uso consapevole e critico del linguaggio nei prompt possa diventare un potente mezzo espressivo per ideare e realizzare contenuti visivi originali
L'Arte del prompt nell'era dell'IA generativa per immagini
DESSI', ANNA
2024/2025
Abstract
This thesis analyzes the role of the prompt in the use of generative artificial intelligence for image creation, with a particular focus on text-to-image models. After an overview of the origins of AI, the work explores the limitations of these systems, especially algorithmic bias, visual hallucinations, and issues related to copyright. A central part of the research is devoted to prompt engineering, understood as a strategic methodology for guiding generative output through accurate and effective textual design. The thesis aims to identify techniques and approaches that enhance the quality of generated results by fostering a creative interaction between user and machine. The final section presents a practical case study: the design of an illustrated calendar developed in collaboration with the real estate agency “Tetto Rosso”, using a workflow that integrated tools such as ChatGPT, Midjourney, and Photoshop. This project demonstrates how the conscious and critical use of language in prompts can become a powerful expressive tool for designing and producing original visual content.| File | Dimensione | Formato | |
|---|---|---|---|
|
Dessi_Anna.pdf
Accesso riservato
Dimensione
1.75 MB
Formato
Adobe PDF
|
1.75 MB | Adobe PDF |
The text of this website © Università degli studi di Padova. Full Text are published under a non-exclusive license. Metadata are under a CC0 License
https://hdl.handle.net/20.500.12608/87825