The distinction between distributive and non-distributive profiles figures prominently in current evaluations of the ethical and epistemological risks that are associated with automated profiling practices. The diagnosis that non-distributive profiles may coincidentally situate an individual in the wrong category is often perceived as the central shortcoming of such profiles. According to this diagnosis, most risks can be retraced to the use of non-universal generalisations and various other statistical associations. This article develops a top-down analysis of non-distributive profiles in which this (...) fallibility of non-distributive profiles is no longer central. Instead, it focuses on how profiling creates various asymmetries between an individual data-subject and a profiler. The emergence of informational, interest, and perspectival asymmetries between data-subject and profiler explains how non-distributive profiles weaken the epistemic position of a profiled individual. This alternative analysis provides a more balanced assessment of the epistemic risks associated with non-distributive profiles. (shrink)

In this paper we make the case for the emergence of novel kind of bias with the use of algorithmic decision-making systems. We argue that the distinctive generative process of feature creation, characteristic of machine learning (ML), contorts feature parameters in ways that can lead to emerging feature spaces that encode novel algorithmic bias involving already marginalized groups. We term this bias _assembled bias._ Moreover, assembled biases are distinct from the much-discussed algorithmic bias, both in source (training data versus feature (...) creation) and in content (mimics of extant societal bias versus reconfigured categories). As such, this problem is distinct from issues arising from bias-encoding training feature sets or proxy features. Assembled bias is not epistemically transparent in source or content. Hence, when these ML models are used as a basis for decision-making in social contexts, algorithmic fairness concerns are compounded. (shrink)

The emergence of a global industry of digital health platforms operated by Big Tech corporations, and its growing entanglements with academic and pharmaceutical research networks, raise pressing questions on the capacity of current data governance models, regulatory and legal frameworks to safeguard the sustainability of the health research ecosystem. In this article, we direct our attention toward the challenges faced by the European General Data Protection Regulation in regulating the potentially disruptive engagement of Big Tech platforms in health research. The (...) General Data Protection Regulation upholds a rather flexible regime for scientific research through a number of derogations to otherwise stricter data protection requirements, while providing a very broad interpretation of the notion of “scientific research”. Precisely the breadth of these exemptions combined with the ample scope of this notion could provide unintended leeway to the health data processing activities of Big Tech platforms, which have not been immune from carrying out privacy-infringing and socially disruptive practices in the health domain. We thus discuss further finer-grained demarcations to be traced within the broadly construed notion of scientific research, geared to implementing use-based data governance frameworks that distinguish health research activities that should benefit from a facilitated data protection regime from those that should not. We conclude that a “re-purposing” of big data governance approaches in health research is needed if European nations are to promote research activities within a framework of high safeguards for both individual citizens and society. (shrink)

Data science, and the related field of big data, is an emerging discipline involving the analysis of data to solve problems and develop insights. This rapidly growing domain promises many benefits to both consumers and businesses. However, the use of big data analytics can also introduce many ethical concerns, stemming from, for example, the possible loss of privacy or the harming of a sub-category of the population via a classification algorithm. To help address these potential ethical challenges, this paper maps (...) and describes the main ethical themes that were identified via systematic literature review. It then identifies a possible structure to integrate these themes within a data science project, thus helping to provide some structure in the on-going debate with respect to the possible ethical situations that can arise when using data science analytics. (shrink)

The debate on the ethics of privacy has been mainly dominated by Western perspectives, to the exclusion of broader ethical theories and socio-cultural perspectives. This imbalance carries risks; transplanted ethical norms and values can collide with those of the communities in which they are deployed. The consequent homogenization might also represent a missed opportunity to enrich and develop the current paradigm of privacy protection so as to effectively face new technological challenges. This article introduces and discusses the sub-Saharan philosophy of (...) Ubuntu and argues how its conception of the self helps to reinterpret some of the emerging issues revolving digital information technologies. To begin with, a general overview of the debate on the ethics of privacy is provided by distinguishing between individual and relational privacy. Also, the challenges of 'group privacy' are discussed. Then, we introduce basic principles of Ubuntu focusing on how these may have affected privacy conceptions and related legal practices. By outlining opportunities and risks of intercultural information ethics, we argue how Ubuntu—similarly to other communitarian moral philosophies—strengthens the development of the concept of relational privacy and, in particular, of group privacy. (shrink)

This paper delves into the nuances of ‘fashion’ in recommender systems and social media analytics, which shape and define an individual’s perception and self-relationality. Its aim is twofold: first, it supports a different perspective on privacy that focuses on the individual’s process of identity construction considering the social and personal aspects of ‘fashion’. Second, it underlines the limitations of computational models in capturing the diverse meaning of ‘fashion’, whereby the algorithmic prediction of user preferences is based on individual conscious and (...) unconscious associations with fashion identity. I test both of these claims in the context of current concerns over the impact of algorithmic personalisation systems on individual autonomy and privacy: creating ‘filter bubbles’, nudging the user beyond their conscious awareness, as well as the inherent bias in algorithmic decision-making. We need an understanding of privacy that sustains the inherent reduction of fashion identity to literal attributes and protects individual autonomy in shaping algorithmic approximations of the self. (shrink)

In the digital age individuals face key choices about whether and how to share intimate details of their lives, “images of the body, biological data in general and diagnostic information” (Pyrrho,...

Data analytics and data-driven approaches in Machine Learning are now among the most hailed computing technologies in many industrial domains. One major application is predictive analytics, which is used to predict sensitive attributes, future behavior, or cost, risk and utility functions associated with target groups or individuals based on large sets of behavioral and usage data. This paper stresses the severe ethical and data protection implications of predictive analytics if it is used to predict sensitive information about single individuals or (...) treat individuals differently based on the data many unrelated individuals provided. To tackle these concerns in an applied ethics, first, the paper introduces the concept of “predictive privacy” to formulate an ethical principle protecting individuals and groups against differential treatment based on Machine Learning and Big Data analytics. Secondly, it analyses the typical data processing cycle of predictive systems to provide a step-by-step discussion of ethical implications, locating occurrences of predictive privacy violations. Thirdly, the paper sheds light on what is qualitatively new in the way predictive analytics challenges ethical principles such as human dignity and the (liberal) notion of individual privacy. These new challenges arise when predictive systems transform statistical inferences, which provide knowledge about the cohort of training data donors, into individual predictions, thereby crossing what I call the “prediction gap”. Finally, the paper summarizes that data protection in the age of predictive analytics is a collective matter as we face situations where an individual’s (or group’s) privacy is violated using dataotherindividuals provide about themselves, possibly even anonymously. (shrink)

This paper sets out the notion of a current “biopolitical turn of digital capitalism” resulting from the increasing deployment of AI and data analytics technologies in the public sector. With applications of AI-based automated decisions currently shifting from the domain of business to customer (B2C) relations to government to citizen (G2C) relations, a new form of governance arises that operates through “algorithmic social selection”. Moreover, the paper describes how the ethics of AI is at an impasse concerning these larger societal (...) and socioeconomic trends and calls for an ethics of AI that includes, and acts in close alliance with, social and political philosophy. As an example, the problem of Predictive Analytics is debated to make the point that data-driven AI (Machine Learning) is currently one of the main ethical challenges in the ethics of AI. (shrink)

By the early 2020s, emotional artificial intelligence will become increasingly present in everyday objects and practices such as assistants, cars, games, mobile phones, wearables, toys, marketing, insurance, policing, education and border controls. There is also keen interest in using these technologies to regulate and optimize the emotional experiences of spaces, such as workplaces, hospitals, prisons, classrooms, travel infrastructures, restaurants, retail and chain stores. Developers frequently claim that their applications do not identify people. Taking the claim at face value, this paper (...) asks, what are the privacy implications of emotional AI practices that do not identify individuals? To investigate privacy perspectives on soft non-identifying emotional AI, the paper draws upon the following: over 100 interviews with the emotion detection industry, legal community, policy-makers, regulators and NGOs interested in privacy; a workshop with stakeholders to design ethical codes for using data about emotions; a UK survey of 2068 citizens on feelings about emotion capture technologies. It finds a weak consensus among social stakeholders on the need for privacy, this driven by different interests and motivations. Given this weak consensus, it concludes that there exists a limited window of opportunity to societally agree principles of practice regarding privacy and the use of data about emotions. (shrink)

Purpose Privacy has been understood as about one’s own information, information that is not one’s own is not typically considered with regards to an individual’s privacy. This paper aims to draw attention to this issue for conceptualizing privacy when one’s privacy is breached by others’ information. Design/methodology/approach To illustrate the issue that others' information can breach one's own privacy, this paper uses real-world applications of forensic genealogy and recommender systems to motivate the discussion. Findings In both forensic genealogy and recommender (...) systems, the individual’s privacy is breached by information that is not one’s own. The information that breached one’s privacy, by its nature, is beyond the scope of an individual, which is a phenomenon that has already been captured by emerging discussions about group privacy. This paper further argues that the underlying issue reflected by the examples of forensic genealogy is an extreme case even under the consideration of group privacy. This is because, unlike recommender systems that rely on large amounts of data to make inferences about an individual, forensic genealogy exposes one’s identity by using only one other individual’s information. This paper echoes existing discussions that this peculiar situation where others’ information breaches one’s own privacy reveals the problematic nature of conceptualizing privacy relying only on individualistic assumptions. Moreover, this paper suggests a relational perspective as an alternative for theorizing privacy. Originality/value This situation that others’ information breached one’s own privacy calls for an updated understanding of not only privacy but also the relationship between the person and their information. Privacy scholars need to renew their ethical language and vocabularies to properly understand the issue, which recent privacy conceptualizations are already doing (e.g. group privacy). (shrink)

Personal data is highly vulnerable to security exploits, spurring moves to lock it down through encryption, to cryptographically ‘cloud’ it. But personal data is also highly valuable to corporations and states, triggering moves to unlock its insights by relocating it in the cloud. We characterise this twinned condition as ‘clouded data’. Clouded data constructs a political and technological notion of privacy that operates through the intersection of corporate power, computational resources and the ability to obfuscate, gain insights from and valorise (...) a dependency between public and private. First, we survey prominent clouded data approaches, suggesting their particular affordances produce distinctive versions of privacy. Next, we perform two notional code-based experiments using synthetic datasets. In the field of health, we submit a patient’s blood pressure to a notional cloud-based diagnostics service; in education, we construct a student survey that enables aggregate reporting without individual identification. We argue that these technical affordances legitimate new political claims to capture and commodify personal data. The final section broadens the discussion to consider the political force of clouded data and its reconstitution of traditional notions such as the public and the private. (shrink)

This paper aims to provide new insights to debates on group privacy, which can be seen as part of a social turn in privacy scholarship. Research is increasingly showing that the classic individualistic understanding of privacy is insufficient to capture new problems in algorithmic and online contexts. An understanding of privacy as an “interpersonal boundary-control process” (Altman, The environment and social behavior, Brooks and Cole, Monterey, 1975) framing privacy as a social practice necessary to sustain intimate relationships is gaining ground. (...) In this debate, my research is focused on what I refer to as “self-determined groups” which can be defined as groups whose members consciously and willingly perceive themselves as being part of a communicative network. While much attention is given to new forms of algorithmically generated groups, current research on group privacy fails to account for the ways in which self-determined groups are affected by changes brought about by new information technologies. In an explorative case study on self-organized therapy groups, I show how these groups have developed their own approach to privacy protection, functioning on the basis of social practices followed by all participants. This informal approach was effective in pre-digital times, but online, privacy threats have reached a new level extending beyond the scope of a group’s influence. I therefore argue that self-determined sensitive topic groups are left facing what I present as a dilemma: a tension between the seemingly irreconcilable need for connectivity and a low threshold, on the one hand, and the need for privacy and trust, on the other. In light of this dilemma, I argue that we need new sorts of political solutions. (shrink)

In this paper, I examine whether the use of artificial intelligence (AI) and automated decision-making (ADM) aggravates issues of discrimination as has been argued by several authors. For this purpose, I first take up the lively philosophical debate on discrimination and present my own definition of the concept. Equipped with this account, I subsequently review some of the recent literature on the use AI/ADM and discrimination. I explain how my account of discrimination helps to understand that the general claim in (...) view of the aggravation of discrimination is unwarranted. Finally, I argue that the use of AI/ADM can, in fact, increase issues of discrimination, but in a different way than most critics assume: it is due to its epistemic opacity that AI/ADM threatens to undermine our moral deliberation which is essential for reaching a common understanding of what should count as discrimination. As a consequence, it turns out that algorithms may actually help to detect hidden forms of discrimination. (shrink)

The conjunction of wireless computing, ubiquitous Internet access, and the miniaturisation of sensors have opened the door for technological applications that can monitor health and well-being outside of formal healthcare systems. The health-related Internet of Things (H-IoT) increasingly plays a key role in health management by providing real-time tele-monitoring of patients, testing of treatments, actuation of medical devices, and fitness and well-being monitoring. Given its numerous applications and proposed benefits, adoption by medical and social care institutions and consumers may be (...) rapid. However, a host of ethical concerns are also raised that must be addressed. The inherent sensitivity of health-related data being generated and latent risks of Internet-enabled devices pose serious challenges. Users, already in a vulnerable position as patients, face a seemingly impossible task to retain control over their data due to the scale, scope and complexity of systems that create, aggregate, and analyse personal health data. In response, the H-IoT must be designed to be technologically robust and scientifically reliable, while also remaining ethically responsible, trustworthy, and respectful of user rights and interests. To assist developers of the H-IoT, this paper describes nine principles and nine guidelines for ethical design of H-IoT devices and data protocols. (shrink)