Making Sense of Algorithmic Bias in Communities of Knowledge
1University of Bucharest, Romania; 2University of Bucharest, Romania; 3Romanian-American University, Romania
Every day, algorithmically mediated actions take part in the design and redesign of the digital society. Organizations use existing knowledge to identify actionable patterns and create predefined rules that will facilitate solving specific problems, claiming to simplify and optimize human decisions and actions. Algorithms impact various areas of life, changing the social construction of sleep through sleep tracking apps, redefining intimacy through dating apps and platforms, informing policing and legal decisions or mediating work contracts.
Recent critical literature about algorithms has targeted the biases, discrimination and asymmetrical power relationships produced by algorithmically mediated decisions. Algorithms are made by organizations and individuals or groups of individuals belonging to a society and acting upon socially generated data. Thus, algorithms reflect and modify previous social biases. Social monitoring and criticism of an emerging algorithmic society is distributed across multiple communities of practice – from scientific disciplines in technical, social and humanistic fields, to journals, blogs, vlogs, wikis, and other forms of knowledge making and sharing. In this paper we identify types of scientific, professional and lay knowledge referring to algorithmic bias, and we discuss the main types of arguments brought forward in these genres and communities of knowledge.
Virtual Rings and Hate Networks on Facebook
1Dep. of Legal and Social Science - Università di Chieti-Pescara (IT); 2Dep. of Legal and Social Science - Università di Chieti-Pescara (IT)
Online hate, especially as manifested by means of social media, is a phenomenon that has currently been extensively studied and has also been a subject of concern at the governmental level. Its ability to mutate rapidly from one form to another has made it difficult to arrive at a satisfactory definition of this social phenomenon.
Examples of hate speech occupy large areas within the field of Facebook: in fact, parts of the informational content of various kinds of online sources have become nothing less than catalysts of hatred. The aim of the study is to create a typological analysis of hate speech, a profile of online ‘haters’, and an analysis of characteristic features of the ‘reply to comment’ confrontation area. It has, in fact, become evident from research that confrontations between ‘haters’, conducted by means of public messages posted on Facebook, account for a large amount of the dialogues that appear on RtC (reply to comment). These posts sometimes degenerate into veritable ‘virtual rings’ having no connection to the subject originally under discussion.
The present research is a multiple case study concerning itself with three areas – immigration, race relations, and antipolitics – that arouse especially virulent hatred in national public opinion. The analysis is divided into four phases: 1) selection of public posts; 2) analysis of discussion networks for each individual post, and clustering of the most frequent ‘reply to comment’ entries; 3) text mining of language used in each area (immigration, race relations, and antipolitics) in order to identify words and expressions typically used; 4) classification of each individual ring in accordance with the Grounded theory method and by means of NVivo software.
Degrees of Deceptions: Faking of and in the Credential Society
University of Cambridge, United Kingdom
How do you know whether the news in the media are true or faked? The challenge of established modes of verification and the loss of public trust in the media world have gained much public attention in recent years. Far less attention has been paid to a similar trend in the sphere of education, where fake degrees and other credential have gained in importance. Established technologies of credentialisation are facing in an age where an increasing number of people do not work in the country where their degrees had been issued, and where non-degree (micro) credentials are mushrooming as part of life-long learning. The challenge is essentially the results of an intensification, digitalisation and internationalisation of the credential society, to use a term that Randall Collins coined almost four decades ago (Collins 1979).
It is against this backdrop that the paper will explore the different attempts to make use of blockchains as a technology of authentication of credentials and the platforms they create. I will first develop a theoretical framework that draws on political and economic sociology as well as critical media studies and that sheds light on the enabling conditions of authority. The case study will be used to explore the extent to which these new technologies and platforms establish new forms of transnational private authority. The paper will end by discussing the implications for universities and other educational providers and the re-articulation of the complex relationship between knowledge, power, authority and legitimacy in a digital age, that some refer to as new Dark Age (see e.g.Bridle 2018).
'The Medicine World Needs Great Data, and Finland's Got It': a Qualitative Study
1Newcastle University, United Kingdom; 2Helsingin Yliopisto, Suomi
Data about the Finnish people is being touted as a new source of economic growth for Finland. The country is being marketed as a place in which the welfare state has created exceptionally valuable data resources (both from its biobanks and healthcare data), as a byproduct of having created a society apparently without socio-economic disparities, with excellent health care, high levels of education, and tech savvy. Such claims draw from the domain of bioethics to advance the argument that Finland is a more appropriate place for human subject research, particularly in terms of medicine and health, than the global south and other places in the global north in which far greater inequalities exist. The bioethical argument that people in Finland are not vulnerable and their recruitment for health studies not exploitative is a key strategy in this marketing. However, it is also being deployed now for the recruitment of Finns for clinical trials of pharmaceutical products, such as the vaccine trials in Benin that used Finnish people as trial participants. If health data is a resource then, like all publicly generated resources, there are also socio-political implications to its marketisation, in addition to these ethical considerations. This paper draws from recent fieldwork in Finland with health data creators, researchers, and policymakers, exploring the creation of Finnish health data as a resource with value.