POTs: Protective Optimization Technologies

Abstract

Algorithmic fairness aims to address the economic, moral, social, and political impact that digital systems have on populations through solutions that can be applied by service providers. Fairness frameworks do so, in part, by mapping these problems to a narrow definition and assuming the service providers can be trusted to deploy countermeasures. Not surprisingly, these decisions limit fairness frameworks' ability to capture a variety of harms caused by systems. We characterize fairness limitations using concepts from requirements engineering and from social sciences. We show that the focus on algorithms' inputs and outputs misses harms that arise from systems interacting with the world; that the focus on bias and discrimination omits broader harms on populations and their environments; and that relying on service providers excludes scenarios where they are not cooperative or intentionally adversarial. We propose Protective Optimization Technologies (POTs). POTs provide means for affected parties to address the negative impacts of systems in the environment, expanding avenues for political contestation. POTs intervene from outside the system, do not require service providers to cooperate, and can serve to correct, shift, or expose harms that systems impose on populations and their environments. We illustrate the potential and limitations of POTs in two case studies: countering road congestion caused by traffic-beating applications, and recalibrating credit scoring for loan applicants.

Appears in Conference on Fairness, Accountability, and Transparency (FAT* 2020). Bogdan Kulynych and Rebekah Overdorf contributed equally to this work. Version v1/v2 by Seda Gürses, Rebekah Overdorf, and Ero Balsa was presented at HotPETS 2018 and at PiMLAI 2018

References in corpus (12)

• Wild Patterns: Ten Years After the Rise of Adversarial Machine Learning
• Understanding Black-box Predictions via Influence Functions
• Explaining Explanations in AI
• Fairness Beyond Disparate Treatment & Disparate Impact: Learning Classification without Disparate Mistreatment
• Fairness Constraints: Mechanisms for Fair Classification
• Poisoning Attacks to Graph-Based Recommender Systems
• Towards the Science of Security and Privacy in Machine Learning
• Why Do Adversarial Attacks Transfer? Explaining Transferability of Evasion and Poisoning Attacks
• Empirical Risk Minimization under Fairness Constraints
• Adversarial attacks on Copyright Detection Systems
• Questioning the assumptions behind fairness solutions
• Predicting Financial Crime: Augmenting the Predictive Policing Arsenal

Cited by in corpus (7)

• The Ethics of AI in Education
• FATE in AI: Towards Algorithmic Inclusivity and Accessibility
• Epistemic Power in AI Ethics Labor: Legitimizing Located Complaints
• Fairness Implications of Encoding Protected Categorical Attributes
• Optimization's Neglected Normative Commitments
• Transparency, Compliance, And Contestability When Code Is(n't) Law
• Towards a multi-stakeholder value-based assessment framework for algorithmic systems