The Scored Society: Due Process for Automated Predictions

By D. Citron, Frank A. Pasquale. 2014

636 citations recorded by the source corpus.

Subfields: AI Ethics, Fairness & Society

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In AI Ethics, Fairness & Society

Citations in the corpus

FairSight: Visual Analytics for Fairness in Decision Making (2019, background): “These ranking-based decision processes that concern rank-ordering individuals by their likelihood of success or failure have been widely adopted in many domains ranging from policing, recidivism, to job recruiting, and credit rating, which has a great impact on individuals’ lives [8].”
Price of Transparency in Strategic Machine Learning (2016, background): “One approach recently proposed to limit such occurrences of invidious ML is to enforce some kind of ”transparency”, whereby the goal and inner-workings of the algorithm would be made public [2], [3].”
Explaining Explanations in AI (2018, neutral): “…as a type of ‘debugging’; (Kulesza et al., 2015)) and help developers and humans working with a system learn from it (Samek et al., 2017), and enhance the trust between individuals subject to a decision and the system itself (Citron and Pasquale, 2014; Hildebrandt and Koops, 2010; Zarsky, 2013).”
A Nutritional Label for Rankings (2018, background): “While automatic and seemingly objective, rankers can discriminate against individuals and protected groups [1], and exhibit low diversity [9].”
Categories all the way down (2017, background): “Across the contributions is the sense that, as Citron and Pasquale (2014) have suggested, we now live in “scored societies” where increasingly large tracts of social life are subject to these methods, and in an increasingly automated manner.”
Measuring discrimination in algorithmic decision making (2017, background): “Yet decision making by predictive models may discriminate against people, even if the computing process is fair and well-intentioned (Barocas and Selbst 2016; Calders and Zliobaite 2013; Citron and Pasqualle 2014).”
Society-in-the-loop: programming the algorithmic social contract (2017, background): “For instance, many scholars now advocate for expert oversight, by a human operator, over the behavior of ‘killer robots’ or credit scoring algorithms [17].”
The hidden assumptions behind counterfactual explanations and principal reasons (2019, background): “Others have advocated in favor of interactive tools that allow decision subjects to explore the effect of making changes to certain features [6, 18].”
Intersectional Bias in Hate Speech and Abusive Language Datasets (2020, background): “What if the training data, used to detect bias in social media, were itself biased? Then, we would be in a situation where the algorithms that detect hate speech and abusive language, developed to prevent harm to protected groups such as racial minorities and women, exacerbate existing social dispar”
Towards Accountability for Machine Learning Datasets: Practices from Software Engineering and Infrastructure (2020, background): “taset construction [52, 53, 125], as well as problematic consent practices [113]. The lack of accountability to datafied and surveilled populations as well as groups impacted by data-driven decisions [32] has been further critiqued. Taken together, these objections around data raise a serious challe”
Should ChatGPT be Biased? Challenges and Risks of Bias in Large Language Models (2023, background): “Biased models may unfairly penalize certain groups, such as minority communities or individuals with lower socio-economic status, by assigning them lower credit scores or denying them access to loans and ﬁnancial services based on biased assumptions [110, 111, 112].”
Better, Nicer, Clearer, Fairer: A Critical Assessment of the Movement for Ethical Artificial Intelligence and Machine Learning (2019, background): “Spurred in part by advances in machine learning, algorithmic processes and predictive analytics are being applied to domains from criminal justice [1] to consumer finance [2].”
The Profiling Potential of Computer Vision and the Challenge of Computational Empiricism (2019, background): “The ‘online proling’ discussed over the past several decades, typically concerned knowledge production from ‘transaction generated information’ created by interacting with informational environments [41], framing individuals in terms of proxies and scores [32, 45].”