Paper short abstract:

AI systems construct sterreotypical and biased understandings of human experience and anthropology is best suited to challenge their implicit human reductionism. Although the discipline can play a key role in the future of AI ethics research it will need to design more interdisciplinary projects.

Paper long abstract:

With the rise of AI driven technologies, algorithms have replaced paperwork in the construction of social truths (Graeber, 2016); they build truths about who we are, our cultural worlds and our identities. Anthropologists have discussed the implications of big data as meaning construction (see Boellstroff and Mauer, 2015), the powerful discourses of algorithms as culture (Dourish, 2016; Seaver, 2017) or the multiple ways in which people are negotiating with data narratives in everyday life (Barassi, 2017, 2020; Pink et al., 2018; Dourish and Cruz, 2018). However, much more research is needed on the human reductionism implicit to these systems, and the western-centric and biased visions of human nature implicit to these technologies. This paper brings the findings of a three-year ethnographic project on the profiling of children from before birth (Child | Data | Citizen Project, 2016 - 2019) together with the findings of a (non-anthropological) research project aimed at analyzing the discourses around algorithmic profiling in Europe and the critical practices that are emerging against it (The Human Error Project, 2020 – ongoing). The paper will argue that anthropology has a fundamental role to play in the future of AI ethics research and the study of algorithmic profiling. The discipline reminds us that ideas of human nature are not only social and cultural but also political constructions (Sahlins, 2008; Graeber and Sahlins, 2017). Yet to succeed it will need to build projects that are truly interdisciplinary, which consider data-structures, policies, as well as popular media discourses.