Text-to-image (T2I) models enable users worldwide to create high-definition and realistic images through text prompts, where the underrepresentation and potential misinformation of images have raised growing concerns. However, few existing works examine cultural representativeness, especially involving whether the generated content can fairly and accurately reflect global cultures. Combining automated and human methods, we investigate this issue in multiple dimensions quantificationally and conduct a set of evaluations on three prevailing T2I models (DALL-E v2, Stable Diffusion v1.5 and v2.1). Introducing attributes of cultural cluster and subject, we provide a fresh interdisciplinary perspective to bias analysis. The benchmark dataset UCOGC is presented, which encompasses authentic images of unique cultural objects from global clusters. Our results reveal that the culture of a disadvantaged country is prone to be neglected, some specified subjects often present a stereotype or a simple patchwork of elements, and over half of cultural objects are mispresented.
doi.org/10.1145/3613904.3642877
ICT4D has increasingly adopted participation and community involvement to address power imbalances, namely through the figure of the "local". However, this reliance makes assumptions about the nature of the "local" while limiting scrutiny of research approaches. Through a Postcolonial Critical Discourse Analysis, this paper argues that 1) communities are often essentialized in agency-depriving ways, 2) researchers claim substantial discretionary power in representing communities, and 3) participatory approaches are framed as inherently beneficial, obscuring compromises. The analysis suggests participation serves to maintain the status quo. Going forward, ICT4D research should ground claims in evidence, demonstrate community benefits, acknowledge complexities transparently, and question premises that empirical gaps alone justify research. Rather than participation as a panacea, a reflexive ICT4D should scrutinize representational practices and notions of empowerment that may perpetuate inequities.
doi.org/10.1145/3613904.3642748
Direct interaction with cultural heritage (CH) artefacts is frequently unavailable to visitors, offering an opportunity for HCI designers to explore integrating material aspects into digitally-mediated encounters with CH artefacts. We argue that a thorough understanding of the material experiences of CH artefacts can open a novel design space, enabling engaging and meaningful interactions with digital representations. Capitalising on this potential, we present a user study where we systematically explore the material experiences of historic pop-up and movable books. Our analysis identifies five key material qualities to inspire augmentation: fold-ability, slide-ability, tear-ability, age-ability, and print-ability. Highlighting how these material qualities can inspire novel interactions with their digital representations, we present two extended-reality (XR) prototypes of a CH book. With our work, we present HCI designers with a novel approach on designing CH experiences, firmly rooted in materiality, challenging the prevalent paradigms of `technology-driven' or `as-realistic-as-possible' sensory experiences often found in CH-HCI.
doi.org/10.1145/3613904.3642142
This paper analyses practices of data perception and usage, as well as ongoing and envisioned community technology projects carried out by a Masewal Indigenous group in Mexico through their union of cooperatives, Tosepan. Through fieldwork interviews, Masewal participants expressed how they have been appropriating existing technologies for their people’s self-determination. During a workshop, they imagined how diverse knowledges and lived experiences of their worldview, passed down through generations, could be represented and translated into digital practices more broadly. We draw considerations for the HCI community to embrace novel approaches to data and information systems from the community’s concept of Cosmovision, and develop Micro-, Meso- and Macro- lenses within it. Through these lenses, we discuss how technologies could be designed for specific individual practices and connections with nature (Micro-cosmos) while supporting communal actions for autonomy and self-determination (Meso-cosmos), and considering broader worldmaking processes and implications for identity, prosperity, ecology and plural representations (Macro-cosmos).
doi.org/10.1145/3613904.3642598