Many Worlds of AI

Date: 26-28 April 2023

Venue: Jesus College, University of Cambridge

Panel 16: Alternative Practices: Design and Participation

27 April | 3.00 pm | Chair: Dorian Peters | Venue: Bawden Room

Presentation 1: Imagining AI and a prospective Metaverse: A Participatory Speculative Design Case Study from Japan and reflections from Germany

Presenter: Michel Hohendanner and Chiara Ullstein

Abstract: AI development and AI ethics are continually shaped by imaginations of future technology and narratives about socio-technical societies. The emergence and distribution of these narratives is limited with respect to both, how lay citizens and experts differently form and influence these narratives and how narratives from global north are more dominant than from global south. Local perspectives and the expertise of individual everyday life realities represent highly important factors for technology adoption processes, yet they are often neglected in development processes. In our presentation, we introduce a participatory speculative design framework for fostering an accessible AI discourse based on situated knowledge with the aim of contributing to the plurality of technology visions. Building on previous cross-cultural research about speculative perceptions of AI future societies and digital common good imaginations we present how the role and ethics of AI are imagined by Japanese citizens in a workshop on prospective Metaverse societies. Our approach enables sensemaking processes of and through technology design, revealing participants' anticipations of pressing societal problems, imagined problem solving capacities of AI technology, and underlying societal value systems. The resulting visions discuss AI enhanced prospective democratic practice in the light of pressing problems of Japan‘s silver democracy. Furthermore, future social relationships influenced by Japanese VR-culture and AI technology impacting the construction of identity, the preservation of Japanese food culture, and possibilities of natural disaster preparedness are discussed. These results show that the applied methodological approach allows to highlight how situated knowledge informs the imagination of future technologies. We conclude the presentation by illustrating how intercultural discourse can emerge from presenting the speculated futures as discussion starters to people in other countries, here Germany, and how the reflection on local visions can contribute to the creation and appreciation of futures taking into account local contexts.

Author bios: Michel Hohendanner is a research associate at the Munich Center for Digital Sciences and AI at the Munich University of Applied Sciences (MUAS), Germany. He is also a doctoral candidate at the Faculty for Industrial Design at the University of Wuppertal. Previously he was a visiting researcher at KEIO University in Tokyo, Japan, with Prof. Osawa and at the Kyoto Institute of Technology, Japan, with Prof. Mizuno. His research interests include the social impact of increasingly digitalized living environments and the role of design in these contexts.

Chiara Ullstein is a research associate and doctoral candidate at the Chair of Cyber Trust (Department of Informatics) at the Technical University of Munich (TUM). Previously she was a visiting researcher at KEIO University in Tokyo, Japan, with Prof. Osawa, and at the Kyoto Institute of Technology, Japan, with Prof. Mizuno. In her research she investigates citizen perception of AI and citizen participation in AI development and AI regulation. Her current cross-cultural studies focus on the perception of facial processing technologies and future visions of AI and the Metaverse.

Presentation 2: Intercultural AI: Exchange, Dialogue, and Conflict

Presenter: Cornelius Onimisi Adejoro and Tom Yeh.

Abstract: Students from low-resourced language backgrounds, like those from the Global South region, are secluded within the AI system, they need an opportunity to express themselves and put forward their opinion on how best to come up with a more inclusive AI system that allows diverse cultural backgrounds and collaboration. AI systems developed for students from the Global South may not conform to the exact need of the students. In addressing the problem, we, the Imagine AI team based at the University of Colorado Boulder in the USA, collaborated with partners in Nigeria, Brazil, and South Korea to organize a series of writing contests for local high school students in those countries to write about AI. The students wrote short articles and stories about AI to compete for prizes. They wrote in their local languages to express their opinions about AI and make recommendations based on their desires. Our US-based team consists of international students who also came from Nigeria, Brazil, and South Korea to ensure an accurate and unbiased understanding of students’ writings. We analyzed the written stories collected across the countries we consider to be underrepresented within the AI system to get insight into the sentiment they share. Additionally, we surveyed existing applications that allow for collaborative learning and recommended possible areas of improvement. The pilot work we did in Latin America (Brazil and Mexico) was a success and hence a pointer to the viability of the scope of this project. The implications of this study show that students across Africa, Asia, and Latin America appreciate the potential and trends in technological advancements. However, the majority of responses from the students suggest the need for deliberate efforts to be channeled toward incorporating fairness and diversity while upholding ethical values in the AI system. Finally, we recommend the concept of participatory design and usability testing with identified end users in developing these systems.

Author bio: Cornelius Adejoro is a Ph.D. student at the University of Colorado Boulder USA, advised by Prof. Tom Yeh. Mr. Adejero received a Master degree in Computer Science from the University of Nigeria in Nsukka. Mr. Adejero has authored and co-authored several research publications in the area of human-computer interaction, big data, and cybersecurity, such as “Users’ Perception of Social Media Opportunities and Challenges” and “Contemporary View of Human System Integration in System Development Process.” After receiving his Master degree in Computer Science, rather than working in a large technology company, Mr. Adejoro wanted to put his skills and knowledge into serving underprivileged children and decided to take up a teaching job by joining Nigeria-Korea Model School where he taught children in basic school for six years. As a member of Black in AI, Mr. Adejoro is passionate about AI’s potential to transform education in Nigeria and across Sub-Saharan Africa. Given his current research interest, extensive computer science training, and deep AI knowledge, Mr. Adejoro is uniquely qualified to lead the project.

Presentation 3: Responsible Future-Making? Testing Intercultural AI Ethics through the Use of Generative Tools

Presenter: Nikita SW Chiu

Abstract: The recent proliferation of off-the-shelf generative AI tools such as Stable Diffusion had both intrigued and unsettled the artistic and academic world in equal measure. As part of an international consortium project in exploring how traces of the past shape policy development and society's future, the working group on "Traces and Technology" developed a project to examine if generative AI tool can be responsible used to envision a more sustainable and fairer future. Members of the working group (WG) were asked to visualize a more accurate and fairer representation of the world in a more sustainable future. They were aware of the fact that historic/ mainstream representations of the world (cartography), such as the Robinson and the Universal Transverse Mercator System, often reflect geographical distortion, and had long been viewed as associated with colonial emphasis on the "centre" and the "periphery". The group, comprised of members from diverse cultures and backgrounds, ranging from artists, geographers, to aerospace scholars, produced vastly different visualizations. The group will next attempt to synthesize these images with the help of generative AI tool, such as Stable Diffusion, to see if the application of AI and generative software could be used responsibly to synergize diverse visions. The project seeks to put AI ethics to the test, with close monitoring of AI application through participatory decision-making. In doing so, it will not only document diverse conceptions of the "future" by professionals from various sectors and disciplines, but also to make a concrete attempt at synergyzing diverging, and at times contrasting opinion of a fairer and more responsible vision for the future world.

Author bio: Dr Nikita Chiu is Senior Lecturer in Innovation Policy at the University of Exeter. She is also Ad Astra Distinguished Fellow in Robotic and Outer Space Governance at the Space Engineering Research Center at USC. Dr Chiu is a former recipient of the Ernst Mach Grant. Named after the Austrian physicist, the grant enabled the investigation of multilevel governance policy at the city, national, and international (UN) level. Dr Chiu is also a mentor for Space4Women, a network organized by the United Nations Office for Outer Space Affairs(UNOOSA) to support the UN Sustainable Development Goals. She was selected as one of “20 under 35” future leaders to watch by the Space & Satellite Professionals International in 2019 for her work on space sustainability. She is currently working group lead of "Traces and Technology" - part of an international consortium project that examines how historic traces (e.g. data, archives) could inform and shape the governance of future technologies in a more sustainable and responsible manner.