Javascript must be enabled to continue!
Machine learning and power relations
View through CrossRef
AbstractThere has been an increased focus within the AI ethics literature on questions of power, reflected in the ideal of accountability supported by many Responsible AI guidelines. While this recent debate points towards the power asymmetry between those who shape AI systems and those affected by them, the literature lacks normative grounding and misses conceptual clarity on how these power dynamics take shape. In this paper, I develop a workable conceptualization of said power dynamics according to Cristiano Castelfranchi’s conceptual framework of power and argue that end-users depend on a system’s developers and users, because end-users rely on these systems to satisfy their goals, constituting a power asymmetry between developers, users and end-users. I ground my analysis in the neo-republican moral wrong of domination, drawing attention to legitimacy concerns of the power-dependence relation following from the current lack of accountability mechanisms. I illustrate my claims on the basis of a risk-prediction machine learning system, and propose institutional (external auditing) and project-specific solutions (increase contestability through design-for-values approaches) to mitigate domination.
Title: Machine learning and power relations
Description:
AbstractThere has been an increased focus within the AI ethics literature on questions of power, reflected in the ideal of accountability supported by many Responsible AI guidelines.
While this recent debate points towards the power asymmetry between those who shape AI systems and those affected by them, the literature lacks normative grounding and misses conceptual clarity on how these power dynamics take shape.
In this paper, I develop a workable conceptualization of said power dynamics according to Cristiano Castelfranchi’s conceptual framework of power and argue that end-users depend on a system’s developers and users, because end-users rely on these systems to satisfy their goals, constituting a power asymmetry between developers, users and end-users.
I ground my analysis in the neo-republican moral wrong of domination, drawing attention to legitimacy concerns of the power-dependence relation following from the current lack of accountability mechanisms.
I illustrate my claims on the basis of a risk-prediction machine learning system, and propose institutional (external auditing) and project-specific solutions (increase contestability through design-for-values approaches) to mitigate domination.
Related Results
Recreating Prometheus
Recreating Prometheus
Prometheus, chained to a rock, having his liver pecked out by a great bird only for the organ to grow back again each night so that the torture may be repeated afresh the next day ...
A Comparative Study of Some Selected Classifiers on an Imbalanced Dataset for Sentiment Analysis
A Comparative Study of Some Selected Classifiers on an Imbalanced Dataset for Sentiment Analysis
Extracting subjective data from online user generated text documents is made quite easy with the use of sentiment analysis. For a classification task different individual algorithm...
Intercultural Competence Development Among University Students From a Self-Regulated Learning Perspective
Intercultural Competence Development Among University Students From a Self-Regulated Learning Perspective
Abstract. Intercultural competence is defined as a lifelong learning task that can be developed in any intergroup situation. A self-regulated learning model is applied to better un...
Graphic Design for Children with Learning Disabilities Based on the Isaan Mural Painting
Graphic Design for Children with Learning Disabilities Based on the Isaan Mural Painting
The study of 'Graphic design for children with learning disabilities' is a study that delves into learning-disabled children in the Isaan region. The author used the survey to form...
Machine Learning and Data Sciences for Financial Markets
Machine Learning and Data Sciences for Financial Markets
Leveraging the research efforts of more than sixty experts in the area, this book reviews cutting-edge practices in machine learning for financial markets. Instead of seeing machin...
Neutrosophic Hybrid Machine Learning Algorithm for Diabetes Disease Prediction
Neutrosophic Hybrid Machine Learning Algorithm for Diabetes Disease Prediction
Because of its far-reaching effects, diabetes remains a major health problem on a worldwide scale. It's a metabolic illness that causes hyperglycemia and a host of other health iss...
Learning with ANIMA
Learning with ANIMA
The paper develops a semi-formal model of learning which modifies the traditional paradigm of artificial neural networks, implementing deep learning by means of a key insight borro...
PENINGKATAN AKTIVITAS DAN HASIL BELAJAR MELALUI METODE JIGSAW
PENINGKATAN AKTIVITAS DAN HASIL BELAJAR MELALUI METODE JIGSAW
This study aimed to determine the effect of the use of jigsaw cooperative learning on the activeness and learning outcomes of Civics learning in class VIII.1 SMP Negeri 2 Sungayang...