Socially Responsible AI Algorithms: Issues, Purposes, and Challenges

Abstract

In the current era, people and society have grown increasingly reliant on artificial intelligence (AI) technologies. AI has the potential to drive us towards a future in which all of humanity flourishes. It also comes with substantial risks for oppression and calamity. Discussions about whether we should (re)trust AI have repeatedly emerged in recent years and in many quarters, including industry, academia, healthcare, services, and so on. Technologists and AI researchers have a responsibility to develop trustworthy AI systems. They have responded with great effort to design more responsible AI algorithms. However, existing technical solutions are narrow in scope and have been primarily directed towards algorithms for scoring or classification tasks, with an emphasis on fairness and unwanted bias. To build long-lasting trust between AI and human beings, we argue that the key is to think beyond algorithmic fairness and connect major aspects of AI that potentially cause AI's indifferent behavior. In this survey, we provide a systematic framework of Socially Responsible AI Algorithms that aims to examine the subjects of AI indifference and the need for socially responsible AI algorithms, define the objectives, and introduce the means by which we may achieve these objectives. We further discuss how to leverage this framework to improve societal well-being through protection, information, and prevention/mitigation.

arXiv comment45 pages, 8 figures

References in corpus (39)

• A Survey on Explainable Artificial Intelligence (XAI): Towards Medical XAI
• Equality of Opportunity in Supervised Learning
• Understanding Black-box Predictions via Influence Functions
• Word Embeddings Quantify 100 Years of Gender and Ethnic Stereotypes
• Counterfactual Fairness
• Fairness in Machine Learning: A Survey
• Computational fact checking from knowledge networks
• Adversarial Attacks and Defences: A Survey
• Fairness Testing: Testing Software for Discrimination
• Fairness-Aware Ranking in Search & Recommendation Systems with Application to LinkedIn Talent Search
• Invariant Risk Minimization
• Mobile Sensor Data Anonymization
• AI Fairness 360: An Extensible Toolkit for Detecting, Understanding, and Mitigating Unwanted Algorithmic Bias
• Extremal Mechanisms for Local Differential Privacy
• A Survey of Learning Causality with Data: Problems and Methods
• A Convex Framework for Fair Regression
• An Empirical Characterization of Fair Machine Learning For Clinical Risk Prediction
• A Short-term Intervention for Long-term Fairness in the Labor Market
• A Meta-Transfer Objective for Learning to Disentangle Causal Mechanisms
• Causality for Machine Learning
• Identifying and Correcting Label Bias in Machine Learning
• The Grey Hoodie Project: Big Tobacco, Big Tech, and the threat on academic integrity
• Privacy in Deep Learning: A Survey
• Causal Reasoning for Algorithmic Fairness
• Penalizing Unfairness in Binary Classification
• Causal Understanding of Fake News Dissemination on Social Media
• Improving Cyberbully Detection with User Interaction
• The Deconfounded Recommender: A Causal Inference Approach to Recommendation
• Moral Framing and Ideological Bias of News
• How to be Fair and Diverse?
• Generative Counterfactual Introspection for Explainable Deep Learning
• Neural Network Attributions: A Causal Perspective
• Uncertainty as a Form of Transparency: Measuring, Communicating, and Using Uncertainty
• Explaining Deep Learning Models using Causal Inference
• Does Facebook Use Sensitive Data for Advertising Purposes? Worldwide Analysis and GDPR Impact
• A Survey on Causal Inference
• A causal framework for discovering and removing direct and indirect discrimination
• Explaining Visual Models by Causal Attribution
• Disinformation in the Online Information Ecosystem: Detection, Mitigation and Challenges

Cited by in corpus (6)

• Towards Green Automated Machine Learning: Status Quo and Future Directions
• Liability regimes in the age of AI: a use-case driven analysis of the burden of proof
• Responsible AI: Gender bias assessment in emotion recognition
• Undesirable Biases in NLP: Addressing Challenges of Measurement
• An Interpretable Graph-based Mapping of Trustworthy Machine Learning Research
• Causal Learning for Socially Responsible AI