An implicit ambiguity in the field of prediction-based decision-making concerns the relation between the concepts of prediction and decision. Much of the literature in the field tends to blur the boundaries between the two concepts and often simply refers to ‘fair prediction’. In this paper, we point out that a differentiation of these concepts is helpful when trying to implement algorithmic fairness. Even if fairness properties are related to the features of the used prediction model, what is more properly called (...) ‘fair’ or ‘unfair’ is a decision system, not a prediction model. This is because fairness is about the consequences on human lives, created by a decision, not by a prediction. In this paper, we clarify the distinction between the concepts of prediction and decision and show the different ways in which these two elements influence the final fairness properties of a prediction-based decision system. As well as discussing this relationship both from a conceptual and a practical point of view, we propose a framework that enables a better understanding and reasoning of the conceptual logic of creating fairness in prediction-based decision-making. In our framework, we specify different roles, namely the ‘prediction-modeler’ and the ‘decision-maker,’ and the information required from each of them for being able to implement fairness of the system. Our framework allows for deriving distinct responsibilities for both roles and discussing some insights related to ethical and legal requirements. Our contribution is twofold. First, we offer a new perspective shifting the focus from an abstract concept of algorithmic fairness to the concrete context-dependent nature of algorithmic decision-making, where different actors exist, can have different goals, and may act independently. In addition, we provide a conceptual framework that can help structure prediction-based decision problems with respect to fairness issues, identify responsibilities, and implement fairness governance mechanisms in real-world scenarios. (shrink)

Due to the technology of machine learning, new breakthroughs are currently being achieved with constant regularity. By using machine learning techniques, computer applications can be developed and used to solve tasks that have hitherto been assumed not to be solvable by computers. If these achievements consider applications that collect and process personal data, this is typically perceived as a threat to information privacy. This paper aims to discuss applications from both fields of personality and image analysis. These applications are often (...) criticized by reference to the protection of privacy. This paper critically questions this approach. Instead of solely using the concept of privacy to address the risks of machine learning, it is increasingly necessary to consider and implement ethical anti-discrimination concepts, too. In many ways, informational privacy requires individual information control. However, not least because of machine learning technologies, information control has become obsolete. Hence, societies need stronger anti-discrimination tenets to counteract the risks of machine learning. (shrink)

In an analysis of artificially intelligent systems for medical diagnostics and treatment planning we argue that patients should be able to exercise a right to withdraw from AI diagnostics and treatment planning for reasons related to (1) the physician’s role in the patients’ formation of and acting on personal preferences and values, (2) the bias and opacity problem of AI systems, and (3) rational concerns about the future societal effects of introducing AI systems in the health care sector.

Discrimination in decision making is prohibited on many attributes, but often present in historical decisions. Use of such discriminatory historical decision making as training data can perpetuate discrimination, even if the protected attributes are not directly present in the data. This work focuses on discovering discrimination in instances and preventing discrimination in classification. First, we propose a discrimination discovery method based on modeling the probability distribution of a class using Bayesian networks. This measures the effect of a protected attribute in (...) a subset of the dataset using the estimated probability distribution. Second, we propose a classification method that corrects for the discovered discrimination without using protected attributes in the decision process. We evaluate the discrimination discovery and discrimination prevention approaches on two different datasets. The empirical results show that a substantial amount of discrimination identified in instances is prevented in future decisions. (shrink)