The 2021 Yearbook of the Digital Ethics Lab

Preface
	References
Contents
Contributors
The European Legislation on AI: A Brief Analysis of Its Philosophical Approach
	References
Informational Privacy with Chinese Characteristics
	1 Introduction
	2 Exploring China’s ‘Privacy Awakening’
	3 The State-Individual Relationship in Confucian Thought
	4 Individualisation and the ‘Great Self’
	5 Informational Privacy as Relational Obligations
	6 Conclusion
	References
Lessons Learned from Co-governance Approaches – Developing Effective AI Policy in Europe
	1 Introduction
	2 Co-governance as an Approach
	3 Approach and Methodology – Challenges to Regulating AI and Identifying Common Themes and Examples
		3.1 Challenges in AI Governance
		3.2 Identifying Examples
	4 Co-governance Examples from Outside AI – Ideas, Implementation and Challenges
		4.1 Complex Supply Chain Management – The Transparency/Accountability Challenge
		4.2 Building Reliable and Professional Services – Transparency/Accountability/Dynamic Challenge
	5 Co-governance as It Relates to AI
		5.1 Multi-stakeholder Monitoring, Certification and Labeling of Products
		5.2 Professionalization
		5.3 Points of Control (Targeted Regulation with Stakeholder Buy-in)
	6 “Lessons Learned” – Recommendations for AI Governance in the EU
	7 Conclusions
	References
State-Firm Coordination in AI Governance
	1 Introduction
	2 Theoretical Framework: Power and Legitimacy
	3 The Role of States and Technology Firms in AI Governance
		3.1 Evaluating State-Driven AI Governance
		3.2 Evaluating Corporate-Driven AI Governance
		3.3 The Case for a Combined Approach to AI Governance
	4 How to Achieve Responsible AI Governance
		4.1 Updating the State for the Digital Age
		4.2 Rebooting Tech Firms with Purpose
		4.3 Avenues Toward Systemic Change
	5 Conclusion
	References
The Impact of Australia’s News Media Bargaining Code on Journalism, Democracy, and the Battle to Regulate Big Tech
	1 Introduction
	2 Outcompeted
	3 Google: Pre-emptive Dodging
	4 Facebook: Strongarm Tactics
	5 Concerns over the Media Code
	6 Predatory Aggression
	7 Conclusion
	References
App Store Governance: The Implications and Limitations of Duopolistic Dominance
	1 Introduction
	2 App Store Governance: Duopolistic Dominance
	3 Instructive Episodes in App Store Governance
		3.1 Babylon
		3.2 Parler
		3.3 Contact Tracing for Covid-19
		3.4 “Smart Voting”
	4 Conclusions
	References
A Legal Principles-Based Framework for AI Liability Regulation
	1 Introduction
	2 Methodology
	3 Ethical and Legal Principles
	4 A Case Study: How to Derive Legal Principles for AI Regulation
	5 General Aims or Meta-principles of AI Regulation
	6 Innovation and Trust
	7 Policy Recommendations
	8 Conclusions
	References
The New Morality of Debt
	1 “Datafied” Lending
	2 Recasting Regulation
	References
Site of the Living Dead: Clarifying Our Moral Obligations Towards Digital Remains
	1 Introduction
	2 Contextualizing Online Death
	3 Methodology
	4 Analysis
		4.1 LoAArch – A Balancing Act
		4.2 LoAGR – The Near and Forgotten Dead
		4.3 LoACur – Selective Memory
	5 Conclusion
	References
The Statistics of Interpretable Machine Learning
	1 Introduction
	2 Local Linear Approximators
	3 Rule Lists
	4 Case-Based Methods
	5 Variable Importance
	6 Conclusion
	References
Formalising Trade-Offs Beyond Algorithmic Fairness: Lessons from Ethical Philosophy and Welfare Economics
	1 Introduction
	2 Definitions
		2.1 Fairness Metrics
		2.2 Acceptability of Inequalities
			Legally Protected Characteristics
			Effort vs. Circumstances
			Source of Inequality
			Takeaways
	3 Lessons from Ethical Philosophy on (In)Equalities
		3.1 Ethical Subjectivity of Algorithmic Fairness
		3.2 Linking Ethical Philosophy to Algorithmic Fairness
	4 Lessons from Welfare Economics
		4.1 Welfare in Algorithmic Ethics: Beneficence and Nonmaleficence
		4.2 Liberty in Algorithmic Ethics: Autonomy and Explicability
			Autonomy: Liberty
			Autonomy: Forgiveness
			Autonomy: Vulnerability
			Explicability
	5 Proposed Method: Key Ethics Indicators
		5.1 Define Success
		5.2 Identify Sources of Inequality
		5.3 Identify Sources of Bias
		5.4 Design Mitigation Strategies
		5.5 Operationalise Key Ethics Indicators (KEIs), Calculate Trade-Offs Between KEIs
		5.6 Select a Model and Provide Justifications
	6 Conclusion
	References
Ethics Auditing Framework for Trustworthy AI: Lessons from the IT Audit Literature
	1 Introduction
	2 Governance, Assurance and Risk
	3 Results
		3.1 Governance
		3.2 Assurance
			Agile Methodologies and Safety-Critical Systems
		3.3 Risk
	4 Further Considerations for EA of AI
		4.1 COBIT as a Structural Analogue for an EA of AI Framework
		4.2 Accounting for Agile
		4.3 Risk Thresholds for Ethics Audits?
	5 Conclusion
	Appendices
		Appendix 1: Methodology
		Appendix 2: AGILE and Safety-Critical Systems
			Scrum Methodology
			R-Scrum Methodology
		Appendix 3: Ethics and Risk Mapping
			CITYCoP Ethics Compliance Process
			CITYCoP Ethics Compliance Matrix
	References
Ethics Auditing: Lessons from Business Ethics for Ethics Auditing of AI
	1 Introduction
	2 Definitions, Purposes, and Motivations
		2.1 Definition: What Is an Ethics Audit?
		2.2 Purpose and Motivations: Why Do Ethics Audits?
	3 Methods: How to Do an Ethics Audit?
	4 Evaluating Ethics Auditing
	5 Lessons for AI Ethics Auditing
	6 Conclusion
	Appendices
		Appendix 1: Methodology
		Appendix 2: General Business Ethics Audit – Radar Chart
		Appendix 3: Ethics Thermometer – Ethical Qualities Matrix
		Appendix 4: Economy for the Common Good Balance Sheet
		Appendix 5: Improved Ethics Audit – Risk Assessment Model
		Appendix 6: List of Abbreviations
	References
AI Ethics and Policies: Why European Journalism Needs More of Both
	1 Introduction
	2 The Issues Raised by AI in Journalism
		2.1 Algorithmic Accountability
		2.2 Excessive Personalization
		2.3 Lack of Media Pluralism
		2.4 Privacy
		2.5 Copyright
	3 AI’s Contribution to Journalism
		3.1 Newsgathering
		3.2 Story Production
		3.3 Content Distribution
	4 The European Policy Context
	5 Recommendations
	6 Conclusions
	References
Towards Equitable Health Outcomes Using Group Data Rights
	1 Introduction
	2 Health Inequities, Datafication, and Automation
	3 Losing the Forest for the Trees – How Individual Data Rights Leave Marginalized Populations Vulnerable
	4 Underrepresented, Underserved and Fighting Back – Two Case Studies on Group Data Rights
		4.1 Indigenous Data Sovereignty
		4.2 Disease Advocacy Organizations
	5 Who Are Group Data Rights For?
	6 Exercising Group Data Rights – Practical Barriers and Solutions
	7 Limitations
	8 Conclusion
	References
Ethical Principles for Artificial Intelligence in National Defence
	1 Introduction
	2 Methodology
	3 Ethical Challenges of AI for Defence Purposes
		3.1 Sustainment and Support Uses of AI
		3.2 Adversarial and Non-kinetic Uses of AI
		3.3 Adversarial and Kinetic Uses of AI
	4 Ethical Guidelines for the Use of AI
		4.1 Responsible
		4.2 Equitable
		4.3 Traceability
		4.4 Reliable and Governable
	5 Five Ethical Principles for Sustainment and Support and Adversarial and Non-kinetic Uses of AI
		5.1 Justified and Overridable Uses
		5.2 Just and Transparent Systems and Processes
		5.3 Human Moral Responsibility
		5.4 Meaningful Human Control
		5.5 Reliable AI Systems
	6 Conclusion
	References