Methods and Critical Issues Training: Histories of AI and Asia: A Discussion

Friday 27 November 2020, 13:00–15:00 GMT

Special Joint SVE

Co-organizer: Prof. Hallam Stevens (NTU Institute of Science & Technology for Humanity)

Speakers:  Dr Poornima Paidipaty (London School of Economics), Dr Gladys Chong (Hong Kong Baptist University), Dr James Wright (The Alan Turing Institute), Prof. Xiao Liu (McGill University)

This session explores histories and narratives of automation, predictive analysis, and non-human intelligence in Asia. Rather than resort to straightforward diffusionist narratives about the spread of AI from West to East or indulge broad generalizations that suggest that Asian cultures are more accepting of robots because 'Eastern' religions do not differentiate as sharply between 'human' and 'non-human' intelligences, this session aims to contextualize the development of AI, cybernetics, and 'smart' technologies over the past century in China, India, Japan and South Korea in particular.

This event is co-organized with Prof. Hallam Stevens (NTU Institute of Science & Technology for Humanity), the editor of a new Science, Technology & Society special issue on AI in Asia: Histories and Narratives. Abstracts follow.

Speakers

Number in the Postcolonial Imagination

Dr Poornima Paidipaty (London School of Economics)

In recent years, Indian academics, scientists and politicians have proudly claimed that the country will play a leading role in the advancement of AI and its use in the development of smart cities, algorithmic governance and economic regulation. Such visions, of technocratic futures, are not entirely new; in important ways they resemble claims around data-driven development first articulated in the early decades after Indian independence and Partition. In the 1950s, India was at the forefront of developing new techniques in large-scale sampling, which provided much needed data on economic production and household welfare. These efforts, pioneered by PC Mahalanobis and a small group of statisticians at the Indian Statistical Institute, underwrote bold claims about the possibilities of data-driven governance. Mahalanobis saw the science of statistics as an indispensable tool for developing nations, arguing that not only was sampling cost effective, it was more accurate than older, laborious techniques for the complete enumeration of populations, crop yields or consumption rates. This talk will examine the place of numbers in the postcolonial imagination, arguing that the totalizing aspirations of colonial quantification gave way to new, probabilistic exercises, in which the construction of a 'representative' sample was both a technical and a political act of making the constituent parts of the nation visible. In the process, the talk will trace a new conceptual genealogy for contemporary claims about AI and its role in governing the future of India.

Xiaomi and the Promises of the Good Life? Issues of Security and Risk in the Making of the Smart Home in China

Dr Gladys Chong (Hong Kong Baptist University)

Informed by Foucault's governmentality, this paper examines how the making of the smart home operates, within the nexus of security and risk, in fostering the state-sanctioned discourse on 'the good life'. Taking Xiaomi (a renowned smart home technology company) as a case study, I trace how commercial practices formulate issues of security and risk in three smart home products: smart door lock, home surveillance camera, and virtual home assistant. The analysis is structured along three levels of relationships: (a) self and technology, (b) the ontological self, and (c) self, the family, and society, to unpack how issues of security and risk are framed. The analysis demonstrates that Xiaomi corresponds to, as well as constitutes, the dominant positive societal discourses on technology that bring trust, reliability, and security to daily life. It further advances the state-driven technologisation of Chinese society, in which subjects are guided to embrace the positive dimensions of technology for self-actualisation and self-management. The 'intelligent' qualities of these technological applications are, however, discursively framed, and built on scenarios that magnify plausible risks, and make explicit social anxieties and vulnerabilities involving other social subjects; and have thus further eroded the already deteriorating social trust.

Suspect AI: Vibraimage, Emotion Recognition Technology, and Algorithmic Opacity

Dr James Wright (The Alan Turing Institute)

Vibraimage is a digital system that quantifies and classifies a subject's mental and emotional state by analysing video footage of the movements of their head. This system has been – and apparently continues to be – used by a number of local and national police forces, nuclear power station operators, airport security, and hospitals in Russia, China, Japan, and South Korea, and has been deployed at two Olympic Games, a FIFA World Cup, and a G7 Summit to identify suspect individuals. Yet there is no reliable evidence that the numerous and sweeping claims made about vibraimage by its developers are substantiated; indeed, most of these claims seem unprovable due to the intangibility of what is supposedly being measured.

In this talk, I will look at how the developers of vibraimage have tried to legitimate the technology, and how it has acquired the power to penetrate public and private security infrastructure across Russia and Asia. I propose the term 'suspect AI' as a way to describe the growing number of systems like vibraimage that algorithmically classify suspects / non-suspects, yet are themselves deeply suspect. Although vibraimage may be dismissed as pseudoscientific, its use by government agencies and major global technology corporations raises questions about how we define AI and its 'effectiveness', and the role of opacity in legitimising suspect AI systems concerned with the classification of emotions, intentions, and inner worlds.

Cybernetic Past(s) of the Future(s)

Prof. Xiao Liu (McGill University)

Human entanglement with information technologies not only involves the redistribution of the sensible but is also highly charged with sociopolitical tension. What does a history of cybernetics and information technology before China’s internet age tell us about our information present and future(s)? This talk will share the research and reflections around my book Information Fantasies: Precarious Mediation in Postsocialist China (University of Minnesota Press, 2019).