Wischmeyer / Rademacher Regulating Artificial Intelligence

1. Auflage 2019
ISBN: 978-3-030-32361-5
Verlag: Springer International Publishing
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, 391 Seiten

Reihe: Law and Criminology

This book assesses the normative and practical challenges for artificial intelligence (AI) regulation, offers comprehensive information on the laws that currently shape or restrict the design or use of AI, and develops policy recommendations for those areas in which regulation is most urgently needed. By gathering contributions from scholars who are experts in their respective fields of legal research, it demonstrates that AI regulation is not a specialized sub-discipline, but affects the entire legal system and thus concerns all lawyers. Machine learning-based technology, which lies at the heart of what is commonly referred to as AI, is increasingly being employed to make policy and business decisions with broad social impacts, and therefore runs the risk of causing wide-scale damage. At the same time, AI technology is becoming more and more complex and difficult to understand, making it harder to determine whether or not it is being used in accordance with the law. In light of this situation, even tech enthusiasts are calling for stricter regulation of AI. Legislators, too, are stepping in and have begun to pass AI laws, including the prohibition of automated decision-making systems in Article 22 of the General Data Protection Regulation, the New York City AI transparency bill, and the 2017 amendments to the German Cartel Act and German Administrative Procedure Act. While the belief that something needs to be done is widely shared, there is far less clarity about what exactly can or should be done, or what effective regulation might look like.  The book is divided into two major parts, the first of which focuses on features common to most AI systems, and explores how they relate to the legal framework for data-driven technologies, which already exists in the form of (national and supra-national) constitutional law, EU data protection and competition law, and anti-discrimination law. In the second part, the book examines in detail a number of relevant sectors in which AI is increasingly shaping decision-making processes, ranging from the notorious social media and the legal, financial and healthcare industries, to fields like law enforcement and tax law, in which we can observe how regulation by AI is becoming a reality.

Thomas Wischmeyer is an assistant professor for public law and information law at the University of Bielefeld. He has been an academic visitor to several law schools, including Yale Law School and the Jean Monnet Center of New York University School of Law. Thomas Wischmeyer publishes and teaches on European and German constitutional law, administrative law and legal theory. His current research focuses on IT security and the role of law in the information society. Timo Rademacher is Assistant Professor of Public Law and New Technologies at the University of Hannover. He has obtained his degrees - inter alia - from the Universities of Heidelberg (Dr. iur.) and Oxford (MJur). In Hannover, he publishes on and teaches European and German constitutional and information law. His current research aims at the development of a regulatory framework for the use of Big Data analytics by State and EU public bodies.

Autoren/Hrsg.

Wischmeyer, Thomas

Rademacher, Timo

Inhaltsverzeichnis

1;Preface: Good Artificial Intelligence;5
1.1;Mission and Methodology;5
1.2;Artificial Intelligence;6
1.3;Structure and Content;7
1.4;References;9
2;Contents;10
3;Contributors;12
4;Artificial Intelligence as a Challenge for Law and Regulation;14
4.1;1 Fields of Application for Artificial Intelligence;15
4.2;2 Levels of Impact;16
4.3;3 Legal Aspects;18
4.4;4 Modes of Governance;19
4.5;5 Exercising the State´s Enabling Responsibility Through Measures for Good Digital Governance;20
4.5.1;5.1 Displacements in the Responsibility of Public and Private Actors;20
4.5.2;5.2 Innovative Case-law as an Example;22
4.5.3;5.3 System Protection;23
4.5.4;5.4 Systemic Protection;24
4.5.5;5.5 Regulatory Guidelines;24
4.5.6;5.6 Regarding Regulatory Possibilities;25
4.6;6 Obstacles to the Effective Application of Law;27
4.6.1;6.1 Openness to Development;27
4.6.2;6.2 Insignificance of Borders;28
4.6.3;6.3 Lack of Transparency;30
4.6.4;6.4 Concentration of Power;30
4.6.5;6.5 Escaping Legal Constraints;31
4.7;7 Types of Rules and Regulations;31
4.7.1;7.1 Self-structuring;32
4.7.2;7.2 Self-imposed Rules;32
4.7.3;7.3 Company Self-regulation;33
4.7.4;7.4 Regulated Self-regulation;33
4.7.5;7.5 Hybrid Regulation;35
4.7.6;7.6 Regulation by Public Authorities;35
4.7.7;7.7 Techno-regulation;36
4.8;8 Replacement or Supplementation of Legal Measures with Extra-Legal, Particularly Ethical Standards;36
4.9;9 On the Necessity of Transnational Law;37
4.10;References;38
5;Part I: Foundations of Artificial Intelligence Regulation;43
5.1;Artificial Intelligence and the Fundamental Right to Data Protection: Opening the Door for Technological Innovation and Innova...;44
5.1.1;1 The Objective of the Chapter: `Opening the Door´;45
5.1.2;2 Why a Door Needs to Be Opened;46
5.1.2.1;2.1 The Traditional Basic Concept of Data Protection Law: Strict Limits for the Use of AI;46
5.1.2.2;2.2 Constitutionalisation of a Non-constitutional Regulatory Model: The Right to Informational Self-determination in Germany a...;49
5.1.2.2.1;Innovation of the Fundamental Rights by Using a Regulatory Model of Ordinary Law;49
5.1.2.2.2;Distinction Between the Legal Construction and the Theoretical Concept: The Right to Informational Self-determination as an In...;50
5.1.2.2.3;Supra-individualistic Elements in the Jurisprudence of the Federal Constitutional Law;51
5.1.2.3;2.3 Interim Conclusion;52
5.1.3;3 How the Door Can Be Opened;53
5.1.3.1;3.1 Article 8 CFR as an Obligation of the Legislator to Regulate Data Processing, Not as a Right to Informational Self-determi...;54
5.1.3.1.1;Wording and History of Legislation;54
5.1.3.1.2;Structure;55
5.1.3.1.3;Article 8 as a Modern Fundamental Right to Data Protection: Combination of Openness to Innovation and Effective Protection;56
5.1.3.2;3.2 Looking at the CJEU: The Window of Opportunity;58
5.1.4;4 No Door Without a Frame: The Legislator and Legal Scholarship Have a Responsibility;59
5.1.4.1;4.1 Legislator: Framing AI;59
5.1.4.2;4.2 Legal Scholarship: Plea for More Bottom-up Research;61
5.1.5;References;61
5.2;Artificial Intelligence and Autonomy: Self-Determination in the Age of Automated Systems;64
5.2.1;1 Introduction and Practical Applications;65
5.2.1.1;1.1 Health or Life Insurance and Health Status;65
5.2.1.2;1.2 Financial Transactions and Financial Creditworthiness;66
5.2.1.3;1.3 Chinese Social Credit System and Personal Behaviour in General;67
5.2.2;2 Potential Risks from Automated Systems on Self-Determination;68
5.2.2.1;2.1 The Function of Automated Systems;68
5.2.2.2;2.2 Concrete Structures of Risks;69
5.2.2.2.1;Potential to Influence Automated Decisions in General;69
5.2.2.2.2;Approximation and Standardisation;71
5.2.2.2.3;Lack of Comprehensibility;72
5.2.3;3 Legal Framework Conditions: Individual Self-Determination;73
5.2.3.1;3.1 Legally Affected Rights;73
5.2.3.2;3.2 Requirement of Legal Regulation or Personal Responsibility?;74
5.2.4;4 Possibilities for Legal Responses;75
5.2.4.1;4.1 Instruments with Direct Effect;75
5.2.4.1.1;Preliminary Question: The Legal Relevance of Influencing People;75
5.2.4.1.2;Illegal Influences in the Sense of the Unfair Commercial Practices Directive;76
5.2.4.1.3;Randomly Appearing Criteria: Relevance of a Reasonable Expectation Horizon;78
5.2.4.1.4;Factual Criteria: Profiling and Consent;79
5.2.4.2;4.2 Instruments with Indirect Effect;81
5.2.5;References;82
5.3;Artificial Intelligence and Transparency: Opening the Black Box;85
5.3.1;1 Introduction;86
5.3.2;2 Generating Knowledge Through Access to Information;89
5.3.2.1;2.1 Epistemic Constraints;90
5.3.2.2;2.2 Normative Constraints;92
5.3.2.3;2.3 Functional Constraints;95
5.3.3;3 Creating Agency Through Explanations;97
5.3.3.1;3.1 Causal Explanations;98
5.3.3.2;3.2 Counterfactual Explanations;100
5.3.3.3;3.3 Explanations in Context;102
5.3.4;4 The Way Forward;103
5.3.4.1;4.1 Principles;103
5.3.4.2;4.2 Practice;104
5.3.5;References;107
5.4;Artificial Intelligence and Discrimination: Discriminating Against Discriminatory Systems;112
5.4.1;1 Introduction;113
5.4.2;2 Discriminatory Systems?;113
5.4.2.1;2.1 Flawed Data Collection;114
5.4.2.2;2.2 Flawed Data Aggregation;115
5.4.2.3;2.3 Normative Unresponsiveness;115
5.4.3;3 Current Antidiscrimination Law Doctrine and AI;118
5.4.3.1;3.1 From Causality to Correlations;118
5.4.3.1.1;Imputed Causality;119
5.4.3.1.2;Indirect Discrimination;120
5.4.3.2;3.2 Statistics on Statistics;122
5.4.4;4 Towards a Paradigm of Knowledge Creation;124
5.4.4.1;4.1 A Need for Concepts;124
5.4.4.2;4.2 A Need for Facts;125
5.4.4.3;4.3 Statistics in Court;126
5.4.5;5 Conclusion;127
5.4.6;References;128
5.5;Artificial Intelligence and Legal Personality: Introducing ``Teilrechtsfähigkeit´´: A Partial Legal Status Made in Germany;131
5.5.1;1 Introduction;132
5.5.2;2 The `Dual Dilemma´;133
5.5.2.1;2.1 The Case for Agency;133
5.5.2.2;2.2 The First Dilemma;135
5.5.2.3;2.3 The Case for Legal Personality;136
5.5.2.3.1;Contract Formation;137
5.5.2.3.2;Torts;138
5.5.2.4;2.4 The Second Dilemma;140
5.5.3;3 A `Halfway Status´ Made in Germany: Introducing Teilrechtsfähigkeit;141
5.5.3.1;3.1 Teilrechtsfähigkeit;141
5.5.3.2;3.2 Teilrechtsfähigkeit for Intelligent Agents;144
5.5.3.2.1;Contract Formation and Torts;144
5.5.3.2.2;Side Note: Criminal Law;146
5.5.4;4 Of Mice and Machines;147
5.5.5;5 Conclusion;148
5.5.6;References;149
6;Part II: Governance of and Through Artificial Intelligence;151
6.1;Artificial Intelligence and Social Media;152
6.1.1;1 Of Arms Races and Filter Biases;153
6.1.2;2 Phenomenology;156
6.1.2.1;2.1 Key Features of Social Media;156
6.1.2.2;2.2 Social Media AI;157
6.1.3;3 Concepts of Social Media Regulation;158
6.1.4;4 Protective Social Media Regulation;159
6.1.4.1;4.1 Communication Through AI As an Exercise of Fundamental Rights;160
6.1.4.2;4.2 Restricting AI-Based Communication on Social Media Services;162
6.1.4.2.1;Grounds for Restriction;163
6.1.4.2.2;Means of Restriction;165
6.1.4.3;4.3 Responsibilities of Social Media Providers;166
6.1.4.3.1;Responsibility for Unwanted Content (`Output-Responsibility´);167
6.1.4.3.2;Responsibility for Unwanted Blocking (`Input-Responsibility´);170
6.1.5;5 Facilitative Social Media Regulation;173
6.1.5.1;5.1 Purposes and Instruments of Facilitative Regulation of Social Media AI;174
6.1.5.2;5.2 Constitutional Framework of Facilitative Regulation of Social Media AI;175
6.1.6;6 Conclusion;177
6.1.7;References;178
6.2;Artificial Intelligence and Legal Tech: Challenges to the Rule of Law;181
6.2.1;1 Introduction;182
6.2.2;2 Fields of Application;183
6.2.2.1;2.1 Definitions;183
6.2.2.2;2.2 Legal Tech in Private Use: Status Quo and Trends;185
6.2.2.3;2.3 Legal Tech in Public Use: Status Quo and Trends;186
6.2.3;3 Conceptual Differences Between `Law´ and `Code´;188
6.2.3.1;3.1 Application of the Law As a Social Act;188
6.2.3.2;3.2 Code As a Technical Act;189
6.2.3.3;3.3 Misconceptions of `Code´;190
6.2.3.4;3.4 Need for Regulation;191
6.2.4;4 Constitutional Framework;191
6.2.4.1;4.1 Rule of Law and Democracy;191
6.2.4.2;4.2 Right to Privacy;193
6.2.4.3;4.3 Right to Non-discrimination;195
6.2.5;5 Proposals for `Regulation by Design´: A Balancing Act;196
6.2.5.1;5.1 Regulatory Realignment;196
6.2.5.2;5.2 Legal Protection by Design As a Starting Point;197
6.2.5.3;5.3 Regulatory Guidelines;198
6.2.5.4;5.4 The `Human Factor´;200
6.2.5.5;5.5 Balancing Act;200
6.2.6;6 Conclusion;201
6.2.7;References;202
6.3;Artificial Intelligence and Administrative Decisions Under Uncertainty;205
6.3.1;1 Introduction;206
6.3.2;2 Frictions: Machine Learning Versus Administrative Law;207
6.3.2.1;2.1 Investigation;208
6.3.2.2;2.2 Prediction;209
6.3.2.3;2.3 Decision;210
6.3.3;3 Challenges: Machine Learning and Administration Under Uncertainty;211
6.3.3.1;3.1 Generalizability;212
6.3.3.1.1;Neglect of Selection Effects;212
6.3.3.1.2;The Bias-Variance-Tradeoff and Bias Aversion;213
6.3.3.2;3.2 Counterfactual Reasoning;216
6.3.3.2.1;Experimental Administration;216
6.3.3.2.2;Pseudo-Causal Explanations;217
6.3.3.3;3.3 Error Weighting;219
6.3.3.4;3.4 Proportionality;220
6.3.3.5;3.5 Gaming;223
6.3.3.6;3.6 Complexity;224
6.3.4;4 Conclusion;225
6.3.5;Appendix;226
6.3.6;References;227
6.4;Artificial Intelligence and Law Enforcement;230
6.4.1;1 Introduction: Smart Law Enforcement;231
6.4.2;2 The Status Quo of Smart Law Enforcement;232
6.4.2.1;2.1 `Watching´;232
6.4.2.2;2.2 `Reading´;234
6.4.2.3;2.3 `Listening´;236
6.4.2.4;2.4 `Smelling´;236
6.4.2.5;2.5 Everywhere, Always, and Remembering It All;237
6.4.3;3 Constitutional Frameworks;237
6.4.3.1;3.1 Germany;238
6.4.3.2;3.2 European Union;241
6.4.3.3;3.3 United States;243
6.4.4;4 Three Core Issues;246
6.4.4.1;4.1 A Challenge: State-of-the-Art Information or Undemocratic In-Formation?;248
6.4.4.2;4.2 A Chance: Biased Enforcement or Transparent Biases?;251
6.4.4.3;4.3 A Choice: Human or `Perfect´ Enforcement of the Law?;252
6.4.5;5 Conclusion;255
6.4.6;References;255
6.5;Artificial Intelligence and the Financial Markets: Business as Usual?;260
6.5.1;1 AI in the Financial Markets;261
6.5.1.1;1.1 Business-Customer-Relations: `Robo-advisers´;261
6.5.1.2;1.2 Financial Markets and Institutions: `Cyborg Finance´;263
6.5.1.3;1.3 Compliance: `RegTech I´;265
6.5.1.4;1.4 New Players: `FinTech´;266
6.5.2;2 Regulatory Approaches Towards AI in the Financial Markets;267
6.5.2.1;2.1 Global Level;267
6.5.2.2;2.2 European Level;268
6.5.2.3;2.3 National Level;269
6.5.2.3.1;United States;269
6.5.2.3.2;Germany;270
6.5.2.3.3;United Kingdom;271
6.5.3;3 Governance Through and of AI in the Financial Markets;272
6.5.3.1;3.1 Regulation and Supervision Through AI: `RegTech II´;273
6.5.3.1.1;Supervision: `Robocops´;273
6.5.3.1.2;Regulation: `Machine Readable and Executable Rulebook´;273
6.5.3.2;3.2 Vital Aspects of Future AI Governance;274
6.5.3.2.1;Assigning Responsibility;274
6.5.3.2.2;Redefining Systemic Relevance;275
6.5.3.2.3;Insisting on Auditability;275
6.5.3.2.4;Adapting Reporting and Disclosure;275
6.5.3.2.5;Reducing Regulatory Arbitrage;276
6.5.3.2.6;Consumer Protection;276
6.5.3.2.7;Regulatory Independence;277
6.5.4;4 AI and the Financial Markets: To a New Tomorrow;277
6.5.4.1;4.1 First-Mover Advantages;278
6.5.4.2;4.2 Yesterday´s Mistakes;278
6.5.5;References;279
6.6;Artificial Intelligence and Public Governance: Normative Guidelines for Artificial Intelligence in Government and Public Admin...;282
6.6.1;1 A Design Challenge for Government and Administration in Germany;283
6.6.1.1;1.1 The Definition of Artificial Intelligence;283
6.6.1.2;1.2 AI Applications in Government and Administration;285
6.6.2;2 Points of Reference: Between Ethics and Politics;287
6.6.3;3 Guidelines;288
6.6.3.1;3.1 Law;288
6.6.3.1.1;Motivation, Limitation and Design;288
6.6.3.1.2;The New Legislation on Automated Administrative Decisions;290
6.6.3.2;3.2 Technology;291
6.6.3.3;3.3 Organization;291
6.6.3.4;3.4 Strategies;293
6.6.3.5;3.5 Visions;294
6.6.4;4 Outlook;295
6.6.5;References;295
6.7;Artificial Intelligence and Taxation: Risk Management in Fully Automated Taxation Procedures;299
6.7.1;1 Introduction;300
6.7.2;2 Legal Bases;300
6.7.2.1;2.1 Fully Automated Taxation Procedure;300
6.7.2.2;2.2 Risk Management Systems;302
6.7.2.3;2.3 Compatibility of Confidentiality Requirements with Basic Data Protection Regulations;303
6.7.3;3 The Use of AI Within RMS;304
6.7.3.1;3.1 Suitability of AI Within RMS;305
6.7.3.2;3.2 Controlling AI-based RMS;307
6.7.4;4 Conclusion;308
6.7.5;References;308
6.8;Artificial Intelligence and Healthcare: Products and Procedures;311
6.8.1;1 A Brief Case Study;312
6.8.2;2 Legal Framework: European Medical Devices Law in a Nutshell;314
6.8.3;3 Medical Device Term, Section 3 No. 1 MPG;315
6.8.3.1;3.1 Physiological Component;316
6.8.3.2;3.2 Mechanism of Action;317
6.8.3.3;3.3 Intended Purpose;317
6.8.4;4 Market Access Regulation of Software As a Medical Device;318
6.8.4.1;4.1 Categories of Medical Software;318
6.8.4.1.1;Embedded Software;319
6.8.4.1.2;Stand-Alone Software;319
6.8.4.1.3;Accessories on Medical Devices;320
6.8.4.1.4;Software for General Purposes;321
6.8.4.1.5;Subject of Monitoring Under Medical Devices Law;322
6.8.4.2;4.2 Classification Rules;322
6.8.4.2.1;Risk Classes;322
6.8.4.2.2;Risk Dimensions of Software As a Medical Device;324
6.8.4.2.3;In Particular: Classification of Stand-Alone Software;326
6.8.4.3;4.3 Control-Related Process Modules;328
6.8.5;5 Post Market-Entry Surveillance of Software As a Medical Device;330
6.8.5.1;5.1 Statutory Provisions of the MPG;331
6.8.5.2;5.2 Challenging Information Issue;332
6.8.5.2.1;Emergence of the Information Deficit;332
6.8.5.2.2;The Amendments by the MDR As a Solution Approach?;333
6.8.6;6 Outlook;336
6.8.7;References;336
6.9;Artificial Intelligence in Healthcare: Doctors, Patients and Liabilities;340
6.9.1;1 Introduction: Artificial Intelligence in Medicine;341
6.9.1.1;1.1 Applications of Artificial Intelligence in Medicine and Their Benefits and Limitations;341
6.9.1.2;1.2 A Particular Definition for AI in Medicine: Evidence or Reason?;342
6.9.2;2 AI´s Influence on the Status and Role of Physicians and Patients in Their Relation to Each Other in Medical Ethics;345
6.9.2.1;2.1 The Physician-Patient Relationship in Medical Ethics;345
6.9.2.2;2.2 Principles of Medical Ethics and Their Implementation in the Physician-Patient Relationship;346
6.9.2.3;2.3 Challenges Presented by AI in Medical Ethics;347
6.9.3;3 AI´s Influence on the Status and Role of Physicians and Patients in Their Relation to Each Other in Liability Law;350
6.9.3.1;3.1 Issues of Medical Malpractice Law;351
6.9.3.1.1;3.1.1 Informed Consent, Liability for Lack of Information;351
6.9.3.1.2;3.1.2 Liability for Errors in Treatment and the Burden of Proof;353
6.9.3.2;3.2 Issues of Product Liability Law;356
6.9.3.3;3.3 Conclusions Based on Liability Law;357
6.9.4;4 The Status and Role of Physicians in Their Relation to Patients;358
6.9.5;References;361
6.10;Artificial Intelligence and Competition Law;364
6.10.1;1 Introduction;365
6.10.2;2 Artificial Intelligence;366
6.10.2.1;2.1 Definition(s);367
6.10.2.2;2.2 The Current Legal Framework;367
6.10.3;3 Artificial Intelligence and Its Challenges for Competition Law;368
6.10.3.1;3.1 Market Definition;368
6.10.3.1.1;The Competition for AI;368
6.10.3.1.2;The Competition with AI;369
6.10.3.2;3.2 Market Dominance;370
6.10.3.2.1;The Ninth Amendment of the German ARC 2017;370
6.10.3.2.2;`Data Power´;371
6.10.3.2.3;Transparency of Market-Related Information;372
6.10.3.3;3.3 Abuse of Market Power;374
6.10.3.3.1;Standard-Essential Know-How;374
6.10.3.3.2;The Use of AI;375
6.10.3.4;3.4 Prohibition of Cartels and Other Anti-Competitive Agreements;376
6.10.3.4.1;Exchange of Information and Hub-and-Spoke-Scenarios;377
6.10.3.4.2;`Agreements´ Between AI Systems;378
6.10.3.5;3.5 Merger Control;378
6.10.3.6;3.6 Liability;379
6.10.3.7;3.7 Conclusions;381
6.10.3.7.1;Public Enforcement and the Future Legal Framework;381
6.10.3.7.2;Responsibility for Anti-Competitive Behavior;382
6.10.3.7.3;`Duty to Maintain Safeguards´ or `Störerhaftung´ in Competition Law?;383
6.10.3.7.4;Legal Personality of AI Systems;385
6.10.4;References;386