Mehrdad Sedaghat Baghbani In 2015, Google’s Photos app misclassified Black individuals as “gorillas,” exposing critical biases in AI-driven image recognition. This and similar incidents underscore the urgent need to expand AI training datasets beyond Western-centric perspectives and norms. As AI-driven platforms increasingly shape our experiments with typography and visual design, research on non-Roman typography and non-Eurocentric visual cultures is essential to mitigating algorithmic biases and misrepresentation.
This presentation examines how studying complex visual cultures—such as Islamic manuscripts and Iranian poster design—can provide insights into diverse typographic systems. I will analyze and classify Islamic manuscripts from 610 CE to 1900 to explore the historical relationship between typography and image in contemporary Iranian poster design. Additionally, I will reflect on how my “in-between” lived experience as a US-based Iranian design-educator offers an authentic perspective that makes this analysis accessible to a Western audience. By studying historical interplays between type and image in multicultural settings, designers and design educators can develop a structured approach to understanding non-Western visual traditions and their relevance in contemporary design. Such insights can inform the datasets used by AI-driven image generators and design toolkits, fostering more inclusive design practices.
In this talk, I will also argue that the role of designers, design educators, and students extends far beyond that of passive users confronted with rapidly shifting technology. Instead, as creative leaders, they must actively engage with these advancements, enriching AI-driven design practices with historical and cultural knowledge. By seizing this opportunity, designers can shape the future of technology to be more inclusive, ensuring that AI design systems reflect the full spectrum of global visual cultures rather than reinforcing a Western-dominated canon.