Ethics of Artificial Intelligence (The International Library of Ethics, Law and Technology, 41)

This book presents the reader with a comprehensive and structured understanding of the ethics of Artificial Intelligence (AI). It describes the main ethical questions that arise from the use of AI in different areas, as well as the contribution of various academic disciplines such as legal policy, environmental sciences, and philosophy of technology to the study of AI. AI has become ubiquitous and is significantly changing our lives, in many cases, for the better, but it comes with ethical challenges. These challenges include issues with the possibility and consequences of autonomous AI systems, privacy and data protection, the development of a surveillance society, problems with the design of these technologies and inequalities in access to AI technologies. This book offers specialists an instrument to develop a rigorous understanding of the main debates in emerging ethical questions around AI. The book will be of great relevance to experts in applied and technology ethics and to students pursuing degrees in applied ethics and, more specifically, in AI ethics. Acknowledgments Contents Chapter 1: Introduction References Part I: Can an AI System Be Ethical? Chapter 2: Bias and Discrimination in Machine Decision-Making Systems 2.1 Introduction 2.2 Why Machine Failure Is More Serious 2.3 How Machine Learning Works 2.4 What Is Meant by Machine Discrimination 2.4.1 Fairness Through Unawareness 2.4.2 Individual Fairness 2.4.3 Counterfactual Fairness 2.4.4 Group Fairness 2.4.5 Impossibility of Fairness 2.5 What We Are Talking About: Example of Machine Discrimination 2.6 Why Machine Learning Can Discriminate 2.7 How Machine Discrimination Can Be Overcome 2.7.1 Pre-processing for Fairness 2.7.2 In-training for Fairness 2.7.3 Post-processing for Fairness 2.8 Conclusion References Chapter 3: Opacity, Machine Learning and Explainable AI 3.1 Introduction 3.2 Fundamentals of Trustworthy and Explainable Artificial Intelligence 3.3 Dimensions and Strategies for Promoting Explainablity and Interpretability 3.3.1 Dimensions of Explainability and Interpretability 3.3.2 Interpretability Strategies 3.4 Digging Deeper on Counterfactual Explanations 3.4.1 Basics of Counterfactual Explanations 3.4.2 Overview on Techniques for Counterfactual Explanations 3.5 Future Challenges for Achieving Explainable Artificial Intelligence 3.5.1 Multimodal Data Fusion for Improved Explainability 3.5.2 Reliable and Auditable Machine Learning Systems 3.5.3 GPAI Algorithms to Learn to Explain 3.6 Concluding Remarks References Chapter 4: The Moral Status of AI Entities 4.1 Introduction 4.2 Can Machines Be Moral Agents? 4.3 Do We Need a Mind to Attribute Moral Agency? 4.4 The Challenge of Responsibility 4.5 Artificial Moral Patients and Rights 4.6 Relationalist Proposals 4.7 Conclusion References Part II: Ethical Controversies About AI Applications Chapter 5: Ethics of Virtual Assistants 5.1 Introduction 5.2 What Are Virtual Assistants? 5.3 What Ethical Issues Do Virtual Assistants Raise? 5.3.1 Human Agency and Autonomy 5.3.1.1 Manipulation and Undue Influences 5.3.1.2 Cognitive Degeneration and Dependency 5.3.2 Human Obsolescence 5.3.3 Privacy and Data Collection 5.4 Should We Use Virtual Assistants to Improve Ethical Decisions? 5.5 Concluding Remarks References Chapter 6: Ethics of Virtual Reality 6.1 Introduction 6.2 Preliminaries 6.2.1 Prehistory and History 6.2.2 Is It Real? 6.3 My Avatar 6.3.1 What They Reveal About the User 6.3.2 How They Influence the Behaviour of Other Users 6.3.3 How They Influence the User’s Behaviour 6.4 What Is Good 6.5 What Is Bad 6.5.1 Personal Risks 6.5.2 Social Risks 6.6 What Is Weird 6.7 Ethical Issues 6.7.1 Privacy 6.7.2 Ethical Behaviour 6.8 Conclusions References Chapter 7: Ethical Problems of the Use of Deepfakes in the Arts and Culture 7.1 Introduction: What Is a Deepfake? Why Could It Be Dangerous? 7.2 Are Deepfakes Applied to Arts and Culture Harmful? 7.2.1 Encoding-Decoding and GAN Deepfakes 7.2.2 The Moral Limit of Artistic Illusion 7.2.3 Resurrecting Authors 7.2.4 Falsifying Style 7.3 The Limits of Authorship 7.4 Conclusion References Chapter 8: Exploring the Ethics of Interaction with Care Robots 8.1 Introduction 8.2 State of Art 8.3 What Are Care Robots? 8.3.1 Definition 8.3.2 A Bit of History 8.3.3 Taxonomy 8.3.4 Some More Examples of Existing Robots 8.4 Design 8.5 An Ethical Framework for Care Technologies 8.6 Conclusion References Chapter 9: Ethics of Autonomous Weapon Systems 9.1 Introduction 9.2 Autonomous Weapon Systems 9.2.1 Definitions 9.2.2 Examples 9.2.2.1 Sentry Robots: SGR—A1 9.2.2.2 Loitering Munitions with Human in the Loop: Switchblade and Shahed-136 9.2.2.3 Autonomous Loitering Munitions: HARPY 9.2.2.4 Autonomous Cluster Bomb: Sensor Fuzed Weapon (SFW) 9.2.2.5 Hypothetical AWS: SFW + Quadcopter + Image Recognition Capabilities 9.3 Legal Basis 9.4 Main Issues Posed by AWS 9.4.1 Low Bar to Start a Conflict: Jus Ad Bellum 9.4.2 Availability of Enabling Technologies and the Dual Use Problem 9.4.3 Meaningful Human Control 9.4.4 Unpredictability of AWS 9.4.5 Accountability 9.4.6 Human Dignity—Dehumanization of Targets 9.5 Conclusions References Part III: The Need for AI Boundaries Chapter 10: Ethical Principles and Governance for AI 10.1 Intro: Risk and Governance 10.2 AI Risks, Responsibility and Ethical Principles 10.3 Ethical Guidelines and the European Option for AI Governance 10.4 The Artificial Intelligence Regulation in Europe 10.5 AI Governance: Open Questions, Future Paths References EU Legislation and Official Documents Cited Other Resources Mentioned Chapter 11: AI, Sustainability, and Environmental Ethics 11.1 Introduction 11.2 Energy Demands and Environmental Impacts of AI Applications 11.3 What Is Sustainability? 11.4 A Path to Make AI More Sustainable from Environmental Ethics 11.4.1 The Anthropocentric Concern for the Environmental Costs of AI 11.4.2 The Biocentric Concern for the Environmental Costs of AI 11.4.3 The Ecocentric Concern for the Environmental Costs of AI 11.5 Ethical Values for a Sustainable AI 11.6 Conclusions References Chapter 12: The Singularity, Superintelligent Machines, and Mind Uploading: The Technological Future? 12.1 Introduction 12.2 The Advent of the Singularity: Raymond Kurzweil’s Predictions 12.2.1 Is the Singularity Near? 12.2.2 From Moore’s Law to Law of Accelerating Returns 12.3 The Roadmap to Superintelligent Machines 12.3.1 Concerns and Uncertainties 12.3.2 The Future of Superintelligence by Nick Bostrom 12.4 What if We Can Live Forever? Dreams of Digital Immortality 12.4.1 Types of MU: The Analysis of David Chalmers 12.4.2 Will I Still Be Myself in a Virtual World? Problems with Personal Identity 12.5 Conclusions References Technologies are developing faster and their impact is bigger than ever before. Synergies emerge between formerly independent technologies that trigger accelerated and unpredicted effects. Alongside these technological advances new ethical ideas and powerful moral ideologies have appeared which force us to consider the application of these emerging technologies. In attempting to navigate utopian and dystopian visions of the future, it becomes clear that technological progress and its moral quandaries call for new policies and legislative responses. Against this backdrop, this book series from Springer provides a forum for interdisciplinary discussion and normative analysis of emerging technologies that are likely to have a significant impact on the environment, society and/or humanity. These will include, but be no means limited to nanotechnology, neurotechnology, information technology, biotechnology, weapons and security technology, energy technology, and space-based technologies.