Algorithmic fairness through group parities? The case of COMPAS-SAPMOC

Abstract

Machine learning classifiers are increasingly used to inform, or even make, decisions significantly affecting human lives. Fairness concerns have spawned a number of contributions aimed at both identifying and addressing unfairness in algorithmic decision-making. This paper critically discusses the adoption of group-parity criteria (e.g., demographic parity, equality of opportunity, treatment equality) as fairness standards. To this end, we evaluate the use of machine learning methods relative to different steps of the decision-making process: assigning a predictive score, linking a classification to the score, and adopting decisions based on the classification. Throughout our inquiry we use the COMPAS system, complemented by a radical simplification of it (our SAPMOC I and SAPMOC II models), as our running examples. Through these examples, we show how a system that is equally accurate for different groups may fail to comply with group-parity standards, owing to different base rates in the population. We discuss the general properties of the statistics determining the satisfaction of group-parity criteria and levels of accuracy. Using the distinction between scoring, classifying, and deciding, we argue that equalisation of classifications/decisions between groups can be achieved thorough group-dependent thresholding. We discuss contexts in which this approach may be meaningful and useful in pursuing policy objectives. We claim that the implementation of group-parity standards should be left to competent human decision-makers, under appropriate scrutiny, since it involves discretionary value-based political choices. Accordingly, predictive systems should be designed in such a way that relevant policy goals can be transparently implemented. Our paper presents three main contributions: (1) it addresses a complex predictive system through the lens of simplified toy models; (2) it argues for selective policy interventions on the different steps of automated decision-making; (3) it points to the limited significance of statistical notions of fairness to achieve social goals.

The Challenges of Machine Learning: A Critical Review

“…Bias in data is a pervasive issue, as datasets often mirror the biases and inequalities present in society. Historical and societal biases can be ingrained in various domains, including criminal justice (see, for example, the case of COMPAS [44], employment [45], healthcare [46], and finance [47]. Biased data can result from historical discrimination, cultural stereotypes, or systemic inequalities, leading to under-representation or misrepresentation of certain groups.…”

Liars Know They Are Lying: Differentiating Disinformation from Disagreement

“…It turns out that it is mathematically impossible to simultaneously satisfy both forms of fairness-calibration and classification-when the base rates of re-offending differ between groups (Berk et al, 2021;Lagioia et al, 2023). That is, if a greater share of Black people are classified as high risk-which the algorithm does in an unbiased manner-then it necessarily follows that a greater share of Black defendants who do not re-offend will also be mistakenly classified as high risk.…”

Exploring the Nuptial Bond Between Neuroprediction and AI in Criminal Justice: A Theoretical Review

“…Researchers have also used Google TensorFlow to forecast crime hotspots and evaluated three options in the RNN (Zhuang Y, 2017) architecture: precision, accuracy and recall. A comparative study (McClendon L, Meghanathan N, 2015) between violent crime patterns was carried out using the open-source data mining software WEKA 27 . Here, three algorithms, namely, linear regression, additive regression and decision stump, were implemented to determine the efficiency and efficacy of the ML algorithms.…”