A blind spot in AI-powered logo makers

Visual design principles

Bibliographic Data

ID12700331
AuthorsRenato Antonio Bertão (0000-0003-3566-6155, corresponding author), Myeong-Heum Yeoun (Kookmin University), Jaewoo Joo (0000-0002-2808-5056, Kookmin University)
Year2023
Volume24
Issue1
Pages222-250
Publication date2023-04-17
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueVisual Communication (JOURNAL)
Journal identifiersISSN: 1470-3572 • E-ISSN: 1741-3214
PublisherSAGE Publishing (PUBLISHER • US)
DOI10.1177/14703572231155593
OpenAlexW4366188150
LanguageEN
Citations received2
References cited32

Artificial intelligence is already embedded in several digital tools used across design disciplines. Although it offers advantages in automating and facilitating design tasks, this technology has constraints to empowering practitioners. AI systems steadily incorporate machine learning to deliver meaningful designs but fail in critical dimensions such as creativity. Moreover, the intensive use of AI features to provide a design solution – so-called AI design – challenges the boundaries of the design field and designers’ roles. AI-powered logo makers exemplify a horizon where non-designers can access design tools to create a personal or business visual identity. However, in the current context, these online businesses are limited to randomize layout solutions lacking the visual properties a logo requires. This article reports mixed-method research focusing on AI-powered logo makers’ processes and outcomes. We investigated their capability to deliver consistent logo designs and to what extent their algorithms address logo design principles. Initially, our study identified representative visual principles in logo design-related literature. After probing AI-powered logo makers’ features that enable logo creation, we conducted an exploratory experiment to obtain solutions. Finally, we invited logo design experts to evaluate whether three visual principles (proportion, balance and unity) were incorporated into the layouts. The assessment’s results suggest that these AI design tools must calibrate the algorithms to provide solutions that meet expected logo design standards. Even focusing on a particular AI tool and a few visual principles, our research contributes to initial directions for developing algorithms that embody the complex aspects of visual design syntax

Context (archaeology · Human–computer interaction · Logo (programming language · Programming language · Aesthetic Perception and Analysis · Color perception and design · Computer Science · Digital Media and Visual Art · Artificial Intelligence

Tools

Open DOI