SNF Spark ProjectPublished on 09.09.2020

Reimagining Artificial Intelligence - Towards Establishing a Sociological Conception of AI


The current mechanisms that define and drive the development of Artificial Intelligence (AI) have been widely criticized for being tech-oriented and market-led instead of starting from societal needs and challenges. Subsequently, numerous and notable initiatives have been proposed that argue for proactive engagement of various areas of social science in determining how AI should reach our societies. In these proposals, AI projects are still conceptualized within the tech sector while the attempt is to bring knowledge created in social sciences and humanities to the position of informing AI adoption policies.

In this project, we seek to examine a radically different placement for social sciences in the realm of AI. Our broad objective is to capture the constituents and structure of a ​distinctively sociological conception of AI that is not subordinate to or dependent on the technological trends (e.g. deep learning, pervasive sensing), but rather is grounded solely in the imaginaries of ​what humans need AI to be​.

To the interest of clarity and to be able to achieve a concrete outcome, we confine the scope of this project to a specific domain that is highly influenced by the discourses of “smart agenda”– the domain of urbanism. Moreover, we focus our research on the context of European cities. Therefore, the questions that this project seeks to answer are as follows: In the vision of European cities as ​urban sociologists would describe what are the promoted ideas that can be associated with a form of non-human intelligence? For example, the notion of “knowledge-intensive” has been proposed in the context of smart urbanity in contrast to the current “technology-intensive” model (McFarlane and Söderstrom 2017). Can a collection of such notions and their interrelations create a (domain-specific) alternative “conception of AI”? Also, compared to the existing conception of AI what are the main dissimilarities and similarities?

The outcome will provide a new ground for discussions of AI projects in cities. We do not expect the creation of a new conception of AI lead to the creation of new type of AI devices and tools. Rather, by developing such a radically shifted standpoint the main impact is to achieve a distant sight, pronouncing what AI should bring to societies and making visible the nature and nuances of tensions between what can be beneficial to our societies and how AI is directed currently.

The research is conducted by a multidisciplinary team of (senior) researchers:  Hamed S. Alavi (Computer Science, HCI), Jakub Mlynar (Sociology), Farzaneh Bahrami (Urbanisme), Andre Ourednik (Geography), Himanshu Verma (Computer Science, HCI), Nico Mutzner (Sociology).