According to Clifford Geertz, the purpose of anthropology is not to explain culture but to explicate it. That should cause us to rethink our relationship with machine learning. It is, we contend, perfectly possible that machine learning algorithms, which are unable to explain, and could even be unexplainable themselves, can still be of critical use in a process of explication. Thus, we report on an experiment with anthropological AI. From a dataset of 175K Facebook comments, we trained a neural network (...) to predict the emoji reaction associated with a comment and asked a group of human players to compete against the machine. We show that a) the machine can reach the same accuracy as the players, b) it fails in roughly the same ways as the players, and c) easily predictable emoji reactions tend to reflect unambiguous situations where interpretation is easy. We therefore repurpose the failures of the neural network to point us to deeper and more ambiguous situations where interpretation is hard and explication becomes both necessary and interesting. We use this experiment as a point of departure for discussing how experiences from anthropology, and in particular the tension between formalist ethnoscience and interpretive thick description, might contribute to debates about explainable AI. (shrink)

In this paper, we propose the concept of controversing as an approach for engaging citizens in debates around the datafied city and in shaping responsible smart cities that incorporate diverse public values. Controversing addresses the engagement of citizens in discussions about the datafication of urban life by productively deploying controversies around data. Attempts to engage citizens in the smart city frequently involve ‘neutral’ data visualisations aimed at making abstract sociotechnical issues more tangible. In addition, citizens are meant to gather around (...) issues already defined externally by others. Instead, we focus on how people might become engaged and develop the capacity to shape alternative urban futures. We suggest that making controversial apparently less contentious issues in the smart city allows people to identify their own issues, come together temporarily as a public, imagine alternative possibilities and thus develop capacities for action. In this context, controversies can act as agents of change and open up new spaces for participation and action. We develop the notion of controversing as a deliberate strategy of making datafication controversial, and operationalise the term along the dimensions of recontextualisation, meaning-making and agency. We then look at two cases from the mid-sized city of Amersfoort in the Netherlands, first to test the conceptual potential of controversing to expose how frictions shape citizen engagement, and second to analyse how controversing may frame design-oriented methods aimed at involving diverse participants in discussing datafication and defining public values in the datafied smart city. (shrink)

International relations are made up of thick layers of meaning and big streams of data. How can we capture the nuances and scales of increasingly digitalised world politics, taking advantage of the possibilities that come with ‘big data’ and ‘digital methods’ in our discipline of International Relations? What is needed, we argue, is a methodological twin-move of making big data thick and thick data big. Taking diplomacy, one of IR's core practices as our case, we illustrate how anthropological and computational (...) approaches can be merged in IR research. We report from our experiences with the project DIPLOFACE: Diplomatic Face-Work between Confidential Negotiations and Public Display, investigating how digital communication technologies influence both the study and conduct of age-old and traditionally analogue practices of inter-state diplomacy. (shrink)

Sensor-based data are becoming increasingly widespread in social, behavioral and organizational sciences. Far from providing a neutral window on 'reality', sensor-based big-data are highly complex, constructed data sources. Nevertheless, a more systematic approach to the validation of sensors as a method of data collection is lacking, as their use and conceptualization have been spread out across different strands of social-, behavioral- and computer science literature. Further debunking the myth of raw data, the present article argues that, in order to validate (...) sensor-based data, researchers need to take into account the mutual interdependence between types of sensors available on the market, the conceptual (construct) choices made in the research process, and the contextual cues. Sensor-based data in research are usually combined with additional quantitative and qualitative data sources. However, the incompatibility between the highly granular nature of sensor data and the static, a-temporal character of traditional quantitative and qualitative data has not been sufficiently emphasized as a key limiting factor of sensor-based research. It is likely that the failure to consider the basic quality criteria of social science measurement indicators more explicitly may lead to the production of insignificant results, despite the availability of high volume and high-resolution data. The paper concludes with recommendations for designing and conducting mixed methods studies using sensors. (shrink)