The Ethics of Artificial Intelligence: Principles, Challenges, and Opportunities

Copyright
Contents
Preface
Acknowledgements
List of Figures
List of Tables
List of Most Common Abbreviations and Acronyms
Part One Understanding AI
1 Past: The Emergence of AI
	1.0 Summary
	1.1 Introduction: The Digital Revolution and AI
	1.2 Digital’s Cleaving Power: Cutting and Pasting Modernity
	1.3 New Forms of Agency
	1.4 AI: A Research Area in Search of a Definition
	1.5 Conclusion: Ethics, Governance, and Design
2 Present: AI as a New Form of Agency, Not Intelligence
	2.0 Summary
	2.1 Introduction: What Is AI? ‘I Know It When I See It’
	2.2 AI as a Counterfactual
	2.3 The Two Souls of AI: The Engineering and the Cognitive
	2.4 AI: A Successful Divorce Made in the Infosphere
	2.5 The Human Use of Humans as Interfaces
	2.6 Conclusion: Who Will Adapt to Whom?
3 Future: The Foreseeable Development of AI
	3.0 Summary
	3.1 Introduction: Looking into the Seeds of Time
	3.2 Historical, Hybrid, and Synthetic Data
	3.3 Constraining and Constitutive Rules
	3.4 Difficult Problems, Complex Problems, and the Need for Enveloping
	3.5 Generative Models
	3.6 A Future of Design
	3.7 Conclusion: AI and Its Seasons
Part Two Evaluating AI
4 A Unified Framework of Ethical Principles for AI
	4.0 Summary
	4.1 Introduction: Too Many Principles?
	4.2 A Unified Framework of Five Principles for Ethical AI
	4.3 Beneficence: Promoting Well-Being, Preserving Dignity, and Sustaining the Planet
	4.4 Nonmaleficence: Privacy, Security, and ‘Capability Caution’
	4.5 Autonomy: The Power to ‘Decide to Decide’
	4.6 Justice: Promoting Prosperity, Preserving Solidarity, Avoiding Unfairness
	4.7 Explicability: Enabling the Other Principles through Intelligibility and Accountability
	4.8 A Synoptic View
	4.9 AI Ethics: Whence and for Whom?
	4.10 Conclusion: From Principles to Practices
5 From Principles to Practices: The Risks of Being Unethical
	5.0 Summary
	5.1 Introduction: Risky Translations
	5.2 Ethics Shopping
	5.3 Ethics Bluewashing
	5.4 Ethics Lobbying
	5.5 Ethics Dumping
	5.6 Ethics Shirking
	5.7 Conclusion: The Importance of Knowing Better
6 Soft Ethics and the Governance of AI
	6.0 Summary
	6.1 Introduction: From Digital Innovation to the Governance of the Digital
	6.2 Ethics, Regulation, and Governance
	6.3 Compliance: Necessary but Insufficient
	6.4 Hard and Soft Ethics
	6.5 Soft Ethics as an Ethical Framework
	6.6 Ethical Impact Analysis
	6.7 Digital Preferability and the Normative Cascade
	6.8 Digital Ethics’ Dual Advantage
	6.9 Conclusion: Ethics as a Strategy
7 Mapping the Ethics of Algorithms
	7.0 Summary
	7.1 Introduction: A Working Definition of Algorithm
	7.2 Map of the Ethics of Algorithms
	7.3 Inconclusive Evidence Leading to Unjustified Actions
	7.4 Inscrutable Evidence Leading to Opacity
	7.5 Misguided Evidence Leading to Unwanted Bias
	7.6 Unfair Outcomes Leading to Discrimination
	7.7 Transformative Effects Leading to Challenges for Autonomy and Informational Privacy
	7.8 Traceability Leading to Moral Responsibility
	7.9 Conclusion: The Good and Evil Use of Algorithms
8 Bad Practices: The Misuse of AI for Social Evil
	8.0 Summary
	8.1 Introduction: The Criminal Use of AI
	8.2 Concerns
	8.3 Threats
	8.4 Possible Solutions
	8.5 Future Developments
	8.6 Conclusion: From Evil Uses of AI to Socially Good AI
9 Good Practices: The Proper Use of AI for Social Good
	9.0 Summary
	9.1 Introduction: The Idea of AI for Social Good
	9.2 A Definition of AI4SG
	9.3 Seven Essential Factors for Successful AI4SG
	9.4 Conclusion: Balancing Factors for AI4SG
10 How to Deliver a Good AI Society: Some Recommendations
	10.0 Summary
	10.1 Introduction: Four Ways of Delivering a Good AI Society
	10.2 Who We Can Become: Enabling Human Self- Realization without Devaluing Human Abilities
	10.3 What We Can Do: Enhancing Human Agency without Removing Human Responsibility
	10.4 What We Can Achieve: Increasing Societal Capabilitieswithout Reducing Human Control
	10.5 How We Can Interact: Cultivating Societal Cohesion without Eroding Human Self- Determination
	10.6 Twenty Recommendations for a Good AI Society
	10.7 Conclusion: The Need for Concrete and Constructive Policies
11 The Gambit: AI Impact on Climate Change
	11.0 Summary
	11.1 Introduction: AI’s Double- Edged Power
	11.2 AI and the EU’s ‘Twin Transitions’
	11.3 AI and Climate Change: Ethical Challenges
	11.4 AI and Climate Change: Digital Carbon Footprint
	11.5 Thirteen Recommendations in Favour of AI against Climate Change
	11.6 Conclusion: A More Sustainable Society and a Healthier Biosphere
12 AI and the UN Sustainable Development Goals
	12.0 Summary
	12.1 Introduction: AI4SG and the UN SDGs
	12.2 Assessing Evidence of AI × SDGs
	12.3 AI to Advance ‘Climate Action’
	12.4 Conclusion: A Research Agenda for AI × SDGs
13 Conclusion: The Green and the Blue
	13.0 Summary
	13.1 Introduction: From the Divorce of Agency and Intelligence to the Marriage of Green and Blue
	13.2 The Role of Philosophy as Conceptual Design
	13.3 Back to ‘the Seeds of Time’
	13.4 Green Collars Needed
	13.5 Conclusion: Humanity as a Beautiful Glitch
References
Index