Index

Nathalie A. Smuha

doi:10.1017/9781009427500.011

Index

abstractions

of algorithms, 27

of data-driven systems, 37–38, 40

for welfare fraud, 37–38

of knowledge-driven systems, 32–33

for welfare fraud, 33

accountability

algorithmic rule by law and, 206, 211–213, 218–221, 226–227

diminished accountability, 218–222

transparency and, 218, 226–227

non-arbitrariness principle and, 173

actio popularis, 142

actor-network theory, 51–52

administrative acts

adoption of, 84

algorithmic regulation and, 78–82

automation of legalism and, 213–218

judicial review principle and, 142, 190

algorithmic regulation and, 190, 195–196

opacity in, 59–61

transparency in, 59–61, 128–129, 131–132, 145–146

administrative justice, 16, 106

agency

algorithmic systems’ impact on, 66–69

in Milgram experiment, 67–69

obedience to authority in, 67

AI. See artificial intelligence

AI Act, 14, 47–49, 264–296

AI Advisory Forum, 273

AI Office. See European AI Office

algorithmic regulation under, 274–282

high-risk requirements, 278–294

algorithmic rule by law, 289–295

algorithmic systems under, 264–296

analysis of, 294–296

conformity assessment, 271–272, 286, 293

database of high-risk systems, 273–274, 282, 284, 287

definition of AI under, 268–270

discretion in, 280–284

European AI Board, 272

enforcement of, 272–274

exceptions under, 270

fundamental rights impact assessment, 274–275, 284, 287–288, 304

future-proofness under, 295

general purpose AI models, 276–278

harmonisation approaches to, 288–291

innovation under, 273, 296

list-based approach, 271–282

market surveillance authorities, 272, 287

objectives of, 266–268

public authorities’ requirements under, 274–276, 287–288

record-keeping requirements under, 284–285

risk-based approach to, 265–266, 270–288

disclosure requirements in, 273–278

high-risk systems, 278–288

minimal-risk systems, 271

prohibited practices, 274–276

rule of law and, 16, 265–268, 288–296

Scientific Panel of Independent Experts, 273

scope of, 265–272

social scoring under, 274–275

standards, 273, 293

techno-supremacy in, 291–294

in New Legislative Framework, 293

transparency obligations, 266, 270–277, 284–285, 291–292, 303–305

deep fakes, 276

Treaty on the Functioning of the European Union and, 267–268

Trustworthy AI, 265–266, 288

AI Advisory Forum, 273

AI Convention, Council of Europe and, 23, 290–291

AI Board. See European AI Board

AI effect, 45

algocracy

algorithmic regulation and, 82–87

algorithmic rule by law and, 212

algorithmic regulation, 1–18, 26–93. See also algorithmic rule by law; algorithmic systems; specific topics

under administrative law, 3–4

under AI Act, 274–288

high-risk requirements, 278–288

algocracy and, 82–87

bias and, 57–59, 68, 89, 181–184

conceptualisation of, 26–95

harms and risks of, 5–6, 151–229

hypernudge by, 50

legal certainty principle and, 164–173

correlations in, 168

in data-driven systems, 168–169

foreseeability and, 165–169

in knowledge-driven systems, 168

means-testing algorithmic systems, 169

legalism and, 157–161, 170, 213–218

non-arbitrariness principle, 173–181

discretion and, 177–181

efficiency and, 174–176

explainability reduction, 176–177

government accountability and, 180

by public authorities, 69–93

bureaucracy, 82–87

discretion by, 78–82

rationale of, 87–93

research approach to, 18–25

hypotheses and questions in, 18–19

objectives in, 18–19

relevance of, 24–25

rule of law and, 6–10, 151–225

collective harm under, 7

individual harm under, 6–7

societal harm under, 7–8

separation of powers principle and, 199–205

citizen surveillance, 202–205

executive branch and, 200–201

privatisation of legal infrastructure, 201–202

public and private power, 201–202

technical aspects of, 26–50

algorithmic rule by law, 15–18, 95, 108–109, 205–225

accountability and, 206, 218–221

diminished, 218–221

external/internal review of, 213, 219–221, 263, 294

transparency and, 218–219, 226

administrative justice and, 16

AI Act and, 291–296

algocracy and, 212

analysis of, 151–229

automation of legalism, 206, 213–218

creative compliance and, 215

discretion and, 214–216

coder supremacy and, 206, 210–213

bureaucratic justice, 211

outsourcing of discretion, 213

public authorities and, 211–212

conceptualisation of, 17

due process and, 16

harm theory and, 205–229

individual justice and, 207–209, 214

legality principle and, 152–164, 210, 227

opacity and, 176–177, 206, 215, 218, 224

primacy of techno-rationality and, 207–210

public authorities and, 211–212, 218

public participation and, 228

systemic vulnerability and, 206, 221–225

constitutional retrogression and, 224, 226

LGBTQ+ persons and, 222

threshold for deficiency, 223–224

Treaty on the Functioning of the European Union and, 222–223

technological management and, 212

algorithmic systems. See also specific topics

abstractions of, 27–41

under AI Act, 264–291

artificial intelligence and, 2–3, 44–50

narrow, 22

benefits of, 4–5, 87

classifications of, 30

consistency of, 166–168

data dependency in, 62–66

data-driven systems, 34–43

errors and, 9, 42, 55–61, 162–164

opacity and, 60–61, 162–164

ethics guidelines for, 46–47, 265–266, 288

in France, 24, 89, 172–173, 278

harms and risks of, 6–10, 151–229

human rights violations and, 15, 22, 24, 206, 213, 220–226, 273–274, 309

knowledge-driven systems, 30–34, 41–43

‘many hands’ problem, 68–69, 181

means-testing, 169

opacity in, 6, 9, 50–51, 59–61, 69, 161–162, 172, 218

in Poland, 166–167, 176, 196–197, 203

proxies in, 62–66, 154–155, 162, 183–184, 222, 284

public authorities and, 64–65

public sector use of, 6

societal aspects of, 50–69

bias, 57–59

human error, 55–57

impact on human agency, 66–69

opacity, 59–61

as socio-technical infrastructure, 51–55, 93

affordances, 52

authoritarian technics, 51–52

democratic technics, 51–52

neutrality of, 52

STIR system, 196–197

in Sweden, 167, 172

systemic effects of, 55, 288

algorithms. See also algorithmic regulation; algorithmic systems

abstractions of, 27–41

classifications of, 30

data-driven systems, 34–41

decision-making by, 28–29

support systems, 29

definition of, 27

as historical concept, 27

inputs and, 27–29, 35–36, 39–40, 54

instructions for, 27–28, 31, 36, 45

knowledge-driven systems and, 30–34, 41–43

outputs and, 27–29, 35–43, 47, 49, 53–54, 62

reasoning and, 30–35, 41–43, 156

recommendation system, 29

Arendt, Hannah, 65, 67, 74–75, 217

on banalisation of evil, 67, 217

on bureaucracy, 74–75

Aristotle, on rule of law, 96–97

artificial intelligence (AI), 44–50

under AI Act, 14–15, 47–50, 264–296

AI Convention, 23, 290–291

AI effect, 45

algorithmic systems and, 2–3, 44–50

narrow AI, 22

bias and, 57–59, 68, 89, 181–184

Coordinated Plan on AI, 5, 91–92

Council of Europe, 23, 290

Council of the European Union and, 48, 264, 272, 276

Dartmouth Workshop and, 44

data-driven systems and, 46–47

definition of, 29, 45–50, 268–270

under AI Act, 268–270

by European Commission, 47–50

by OECD, 49

Delvaux Report, 45–47

ethics guidelines for, 46–47, 265–266, 288

European Commission and, 46–50, 91–92, 263–278, 290–291

European Parliament and, 44–46, 48

general purpose, 276

generative, 22, 41–42, 264, 269

under AI Act, 276–278

High-Level Expert Group on AI, 46–47, 265–266, 270

definition of AI, 46–47

ethics guidelines for trustworthy AI, 46–47, 265–266, 288

policy recommendations, 265

high-risk systems, 268–288, 290, 293

minimal-risk systems, 271

prohibited practices, 274–276

standardisation of, 293–294

traditional, 31

Trustworthy AI, 266–267, 281, 288

artificial neural networks, 37

authoritarian technics, 51–52

authoritarian tendencies

expansion of, 107

liberal democracies and, 107, 131, 220

rule of law and, 9–18

constitutional retrogression and, 12–13

authoritarianism

competitive, 107

in hybrid regimes, 107

rule by law and, 105, 107

autocratic regimes, 105, 107, 152, 292, 310

automation

bias, 68, 178, 199, 285

in General Data Protection Regulation, 259–261

in Law Enforcement Directive, 259–261

of legalism, 206, 213–218

creative compliance and, 215

discretion and, 214–216

Bard, 40–41

Bauman, Zygmunt, 75, 136

Belgium. See also Flanders, Belgium

administrative law in, 134, 140, 142–143, 197, 292

algorithmic regulation and, 53, 162, 172, 203–204, 221, 262

judicial review principle in, 142–143, 197

legality principle in, 142

migration law in, 81, 154

Benzécri, Jean-Paul, 35

Berlin, Isaiah, 225

bias. See also discrimination

artificial intelligence and, 57–59, 68, 89, 181–184

in automation, 68, 178, 199, 285

errors from, 57–59

algorithmic regulation and, 57–59, 68, 89, 181–184

big data, 35

Bing Chat, 40–41

Bingham, Lord, 99, 101, 144

black-box systems, 59–61

bureaucracy

Arendt on, 74–75

Bauman on, 75, 136

development of, 75–76

digital, 84–85

limitations and pitfalls of, 73–77

procedural rationality and, 74, 77

by public authorities, 39–73

rational-legal, 75

street-level, 84

substantive rationality and, 74, 76–77

Weber on, 71–78

bureaucratic justice, 211

CAI. See Council of Europe’s Committee on AI

CAHAI. See Council of Europe’s Ad Hoc Committee on AI

Charter of Fundamental Rights of the European Union (CFREU), 66, 112, 118, 127

equality before the law principle in, 138

non-arbitrariness principle and, 133–134

ChatGPT, 40–41

Child Risk Assessment System, 172–173

citizenship, rule of law and, 95

CJEU. See Court of Justice of the European Union

Code of Hammurabi, 96

collective harm, 7

common good, rule of law and, 103

competitive authoritarianism, 107

compliance

creative, 215

with General Data Protection Regulation, 259–260

with Law Enforcement Directive, 259–260

computers

as information highways, 54

screen-level bureaucracy and, 84

Conditionality Regulation, 237–246

contested, 238–241

against Hungary, 240–241

legal basis, 238–239

multi-annual financial framework (MFF), 239

constitutional retrogression, 12–13

algorithmic rule by law and, 224, 226

conformity assessment

under AI Act, 271–272, 286, 293

New Legislative Framework, 293

notified bodies, 272

standardisation, 273

high-risk systems, 282–286

Coordinated Plan on AI, 91–92

correlations

algorithmic regulation and, 168

in data-driven systems, 36

spurious correlations, 38, 42–43

errors and, 56–57

legal certainty principle and, 168

spurious

in data-driven systems, 38, 42–43

errors and, 56

Council of Europe, 23–24, 96, 104, 110–115, 119, 121, 290–291. See also European Convention for the Protection of Human Rights and Fundamental Freedoms

AI Convention, 23, 290–291

Committee of Ministers, 23, 106, 111, 128, 135, 146

rule of law and, 110–112

Venice Commission, 106–108, 110–111, 119, 127, 133, 140, 144, 160

Council of Europe’s Committee on AI (CAI), 23, 290

Council of Europe’s Ad Hoc Committee on AI (CAHAI), 23

feasibility study and, 195

Council of the European Union, 13, 49, 120, 223, 233–235, 238–239, 241, 243, 264, 268, 272

Court of Justice of the European Union (CJEU)

rule of law cases, 112–113, 117–120

Les Verts case, 113

courts. See specific courts

creative compliance, 215

curse of dimensionality, 43

Dartmouth Workshop, AI and, 44

data

big data, 35

dataism, 62

General Data Protection Regulation, 255–263

personal, 14, 65, 146, 255–263

proxy, 184, 284

data dependency, in algorithmic systems, 62–66

data pollution, 8

Data Retention Directive, 249

data-driven methods and systems, 34–41

abstraction of, 37, 40

algorithmic bias and, 57–59

legal certainty principle and, 168–169

automated profiling and, 162

artificial neural networks and, 37

bias in, 58–59

big data and, 35

as black-box systems, 42, 59–61

correlations in, 36, 65, 168, 196

spurious, 38, 42–43, 56

curse of dimensionality, 43

hypernudge and, 50

Internet of Things and, 35

knowledge-driven systems compared to, 41–43

large language models, 40–41

learning, 34–35

deep, 37, 42

reinforcement, 39

semi-supervised, 38

supervised, 36–92

unsupervised, 37–38

legal certainty principle and, 165–169

legality principle and, 156

limitations of, 42–43

transparency in, 59–60

welfare fraud and, 37–38

abstraction of, 38

dataism, 62

decision-making, 58, 66

discretion and, 78–82

public authorities and, 70–73

algorithmic support systems, 29

deep fakes, 276

deep learning, 37, 42, 47, 59, 61, 177

Degrave, Elise, 141, 162–163

Delvaux Report, 45–47

democracy, democracies and

illiberal tendencies, 105–109

constitutional retrogression and, 12–13

rule of law and, 10, 12–13, 101–105

rule by law, 105–109

algorithmic rule by law, 151–205

in EU, 121–125

executive branch of power in, 8

tyranny of the majority, 106, 301

de Tocqueville on, 106–107

democratic technics, 51–52

Dicey, A. V., 97–98

digital bureaucracy, 84–85

Digital Markets Act, EU, 273

Digital Services Act, EU, 255, 267

diminished accountability, 218–221

disclosure requirements, in AI Act, 276, 277, 284, 287

discretion

in AI Act, 280–284

algorithmic regulation and, 177–181

automation of legalism and, 214–216

non-arbitrariness principle and, 177–181

by public authorities, 78–82

supremacy of coders and, 213

discrimination, 38, 58, 111, 135, 138–140, 183, 188, 192–194, 222, 251, 294

due process, 16

Dutch Childcare Allowance case

equality before the law principle and, 184

legality principle and, 159–161

Dworkin, Ronald, 102

EC. See European Commission

ECHR. See European Convention for the Protection of Human Rights and Fundamental Freedoms

ECJ. See European Court of Justice

Economy and Society (Weber), 71–72

ECtHR. See European Court of Human Rights

EEC. See European Economic Community

eGovernment Action Plan 2016–2020, 91

eGovernment Action Plan 2011–2015, 91

Eichmann, Adolf, 67

EP. See European Parliament

EPPO. See European Public Prosecutor’s Office

Equality Act 2010, U.K., 193

equality before the law principle, 138–140

algorithmic regulation and, 181–190

In Allegheny’s child welfare system, 183, 185

comparability loss, 187–190

Dutch Childcare Allowance case and, 184

Ofqual algorithm, 188

scaled bias risks, 181–184

social inequality influenced by, 184–187

transparency and, 189–190

challenges to, 139–140

in Charter of Fundamental Rights of the European Union, 138

in European Convention on Human Rights, 139

requirements of, 138–139

in Treaty on European Union, 148

errors, 55–57

algorithmic systems and, 55

from bias, 57–59

algorithmic regulation and, 59

in data-driven systems, 58–59

correlations and, 56–57

spurious, 38, 56, 196

opacity and, 59–61

ethics

German Data Ethics Commission, 271

guidelines for Trustworthy AI, 46–47, 265–266, 288

Levinas on, 79

ETIAS. See European Travel Information and Authorisation System

EU. See European Union

Eubanks, Virginia, 158, 174, 185

Euclid, 27

European AI Board, 272

European AI Office, 273, 277, 288

European Anti-Fraud Office (OLAF), 236, 242

European Commission (EC). See also AI Act

artificial intelligence and, 264–294

definitions of, 44

ethics guidelines for, 46

rule of law and, 11, 13, 146

European Convention for the Protection of Human Rights and Fundamental Freedoms (ECHR), 110–112, 115

equality before the law principle in, 139

European Court of Human Rights (ECtHR)

judicial review principle and, 140–142, 191

rule of law case law, 112, 115, 118–120, 139, 141

European Court of Justice (ECJ). See CJEU

EU database of high-risk systems, 273, 282, 284, 287, 291

European Parliament (EP), 13, 49, 120, 233–234, 238–239, 244, 264

European Public Prosecutor’s Office (EPPO), 236, 242

European Travel Information and Authorisation System (ETIAS), 308

European Union (EU). See also legal order; specific countries; specific treaties

algorithmic regulation governance in, 230–297

administrative law and, 49

Council of Europe and, 109–112

Charter of Fundamental Rights of the European Union, 66, 112, 118, 127, 133, 138, 146, 243, 250, 256

Council of the European Union, 13, 49, 120, 223, 233–235, 238–239, 241, 243, 264, 268, 272

digital agenda of the EU, 5, 15, 230, 302

Digital Markets Act, 277

Digital Services Act, 255, 267

eGovernment Action Plan 2016-2020, 91

eGovernment Action Plan 2011-2015, 91

General Data Protection Regulation, 187, 255–263, 275, 280, 292, 305

legality principle in, 128–129

Malmo Declaration on eGovernment, 91

recovery plan in, 5, 25, 92, 243

rule by law in, 3–14, 121–125

rule of law in, 109–125

AI Act and, 264–294

in CJEU cases, 112–113, 117–120

conceptualisation of, 109–125

under Conditionality Regulation, 120, 140, 237–246

Copenhagen criteria for, 113–114

as crisis, 4–10, 121, 144

European Commission, 11, 13, 146

European Parliament and, 13, 49, 120, 233–234, 238–239, 244, 264

fragility of, 121–125

horizontal dimensions of, 114–117

as normative legal framework, 125–148

origins of, 112–114

promotion initiatives for, 109

under Treaty of Amsterdam, 114

under Treaty of Lisbon, 114

vertical dimensions of, 114–117

Les Verts case, 113

Treaty on European Union, 3–14, 110, 114

Treaty on the Functioning of the European Union, 114–115, 117–118, 239

values in, 11, 114, 117, 120–125, 266–267

evil, banality of, 67, 217

executive branch, of power

algorithmic regulation and, 9, 69–93

judicial branch and, 199

in liberal democracies, 2–4, 8

public authorities and, 3–4, 21, 70–71

separation of powers principle and, 147–148, 200–201

facial recognition systems, 58, 193, 275

Feinberg, Joel, 5, 7

Flanders, Belgium, 53

formalism, 214

France

algorithmic systems in, 24, 89, 173

Council of State in, 121–122

fraud. See tax fraud; welfare fraud

Fuller, Lon, 99, 144

fundamental rights. See human rights

fundamental rights impact assessment, 274–275, 284, 287–288, 304

future-proofness, under AI Act, 295

Gallie, W. B., 103

General Data Protection Regulation (GDPR), 187, 255–263, 275, 280, 292, 305

automated decision-making in, 259–261

enforcement of, 292–293

evaluation of, 261–263

legal basis, 257–259

in the Netherlands, 257–258

profiling in, 259

general purpose AI system, 276

general purpose AI model, 269, 271, 273, 276–278

Germany, Constitutional Court in, 121–123

Glasius, Marlies, 10, 16

Grosman, Vasily, 79

Hallstein, Walter, 112

handcrafted knowledge, 34

harm, harms and

AI Act, 264–294

algorithmic rule by law, 205–225

citizen surveillance, 202–205

collective, 7

individual, 6–7

societal, 7–8

systemic, 198

Harm Assessment Risk Tool (HART), 185–187

harmonisation approaches, in AI Act, 288–291

HART. See Harm Assessment Risk Tool

Hart, H. L. A., 102, 136–137, 159, 214, 216

Hayek, Friedrich, 98, 132, 137

High-Level Expert Group on AI, 46–47, 265–266, 270

Hildebrandt, Mireille, 4, 150, 153, 178, 221

HMRC. See Revenue and Customs service

Hobbes, Thomas, 1–2

Horizon, 57. See also Post Office scandal

human rights. See also fundamental rights

AI Act and, 265–268

under Charter of Fundamental Rights of the European Union, 66, 112, 118, 127, 133, 138, 146, 243, 250, 256

European Convention for the Protection of Human Rights and Fundamental Freedoms, 110–112, 115

legality and, 126

standardisation and, 127, 273–274, 293

Human Rights Watch, 204–205

human-chosen opacity, 61

Hungary

LGBTQ+ rights in, 123–124

rule of law in, 123, 222, 234

Huq, Aziz, 122–123, 124–125, 224–226, 260

hybrid regimes, 107

Hypereides, 97

Idaho, algorithmic systems in, 55–56, 157

illiberal tendencies and practices

constitutional retrogression and, 12–13

definition of, 10

individual harm, 6–7

individual justice, 207–209, 214

infrastructure, socio-technical, 51–55

infringement actions, 246–255

under Treaty on European Union, 246–247

under Treaty on the Functioning of the European Union, 246–255

innovation, under AI Act, 273, 296

intelligence. See artificial intelligence

Internet of Things (IoT), data-driven systems and, 35

Israel, rule of law crisis in, 12

Jasanoff, Sheila, 129

judicial branch, 21, 145, 147, 220, 228

separation of powers principle and, 145

judicial review principle, 140–144

algorithmic regulation and, 190–199

in ECtHR cases, 191

informational limitations in, 191–195

jurisdictional limitations in, 199

remedy access and, 195–198

systemic review and, 198–199

challenges to, 142–144

requirements of, 140–142

judiciary. See judicial branch; specific courts

justice. See administrative justice; bureaucratic justice; individual justice

al-Khwārizmī, Muḥammad ibn Mūsā, 27

knowledge manipulation, 30–31

knowledge representation, 30–31

knowledge-driven methods and systems, 30–34

abstraction of, 32

for welfare fraud, 33

data-driven systems compared to, 41–43

handcrafted knowledge, 34

reasoning and, 30–32

Krygier, Martin, 99–100

labelling, 36–39, 275

large language models (LLMs), 40–41

Law Enforcement Directive (LED), 255–263, 305

lawfulness. See legality

learning

through data-driven systems, 34–41

deep learning, 37, 42, 59

reinforcement learning, 39

semi-supervised learning, 38

supervised learning, 36–37

unsupervised learning, 37–38

LED. See Law Enforcement Directive

Lefort, Claude, 72, 75–76

legal certainty principle, 131–133

algorithmic regulation and, 164–173

means-testing algorithmic systems, 169

challenges to, 132–133

requirements of, 131–132

legal order, in EU. See also AI Act; specific topics; specific treaties

as autonomous, 10–11, 112–116

competences in, 231–232

Conditionality Regulation and, 237–246

Article 4, 242–243, 245

Data Retention Directive, 249

equality before the law principle, 250

European Anti-Fraud Office, 236, 242

European Parliament and, 13, 49, 120, 233–234, 238–239, 244, 264

European Public Prosecutor’s Office, 236, 242

infringement actions, 246–255

regulation of personal data, 255–263

rule of law in, 109–125

Treaty on European Union and,

Article 2, 11, 127, 138, 232–237, 251–252

Article 7, 223, 233–237, 239–240, 252–255

Treaty on the Functioning of the European Union and, 239–240, 246–247, 254–255

legalism

algorithmic regulation and, 157–161, 170

automation of, 213–218

administrative acts and, 217–218

creative compliance and, 215

discretion and, 214–216

definition of, 158

legality principle and, 157–161

legality, principle of, 126–131

algorithmic regulation and, 152–164

in Belgium, 162

codification of legal rules, 153–157

data-driven systems and, 156

Dutch child allowance case, 159–161

legalism and, 157–161

proportionality tests, 160

transparency and, 161–164, 171–173

in U.K., 163–164

Council of Europe and, 128–129

public participation in, 129

rule of law and, 126–131

challenges to, 129–131

requirements of, 126–129

transparency and, 128–129

Treaty on European Union and, 127

legislative branch, of power

algorithmic regulation and, 220

public authorities and, 70, 145–147

separation of powers principle and, 145, 147–148

Lessig, Lawrence, 3

Leviathan

administrative state and, 98

definition of, 2

symbolic meaning of, 1–2

Leviathan (Hobbes), 1–6

characterisation of, 2

social contract in, 1–2

Levinas, Emmanuel, 79, 214, 225

LGBTQ+ populations

algorithmic rule by law and, 222

rights for, 123–124

Life and Fate (Grosman), 79

Llama, 40–41

LLMs. See large language models

Loughlin, Martin, 97, 103, 109

Lovelace, Ada, 27

Malmo Declaration on eGovernment, 91

manipulation, 274

‘many hands’ problem, 68–69, 181

means-testing algorithmic systems, 169

market surveillance authorities in AI Act, 272, 287

Milgram, Stanley, 67–69, 180

obedience to authority in, 67

mistakes. See errors

Modernity and the Holocaust (Bauman), 75

The Morality of Law (Fuller), 99

Moses, Robert, 52

Mumford, Lewis, 51–52. See also infrastructure

mutual influencing processes, in algorithmic systems, 51

naturalism, rule of law and, 101–102

the Netherlands

General Data Protection Regulation violation in, 257–258

welfare fraud in, 187

Dutch Childcare Allowance case, 108, 159, 184, 198, 262

New Legislative Framework (NLF), 293

New Public Management movement (NPM), 90–91, 179

NLF. See New Legislative Framework

non-arbitrariness principle, 133–138

algorithmic regulation and, 173–181

discretion and, 177–181

efficiency of, 174–176

explainability reduction, 176–177

government accountability and, 180

challenges to, 135–138

Charter of Fundamental Rights of the European Union and, 133–134

notifying authorities, in AI Act, 272

public authorities and, 133–138

public interest and, 137–138

requirements of, 133–135

NPM. See New Public Management movement

OECD. See Organisation for Economic Co-operation and Development

OLAF. See European Anti-Fraud Office

opacity

algorithmic rule by law and, 218–221

errors and, 59–61

in algorithmic systems, 59–62

human-chosen opacity, 60

Organisation for Economic Co-operation and Development (OECD)

artificial intelligence defined by, 49

personal data. See also General Data Protection Regulation

definition of, 256

EU regulation of, 255–263

under Law Enforcement Directive, 255–263

plurality of values, 76

Poland

algorithmic systems in, 166–167, 196

Constitutional Court in, 121–123

LGBTQ+ rights in, 123–124

STIR system in, 196–197

policy recommendations, 302–309

for algorithmic rule by law, 225–229, 302–304

for future research, 307–309

High-Level Expert Group on AI, 265

legal safeguards, 304–306

Politics (Aristotle), 96

positivism, 101–102

Post Office scandal, in U.K., 57, 198

power. See also separation of powers principle

algorithmic regulation and, 199–204

public and private, 146–147, 201–202

principle of lawfulness. See legality

principle of legality. See legality

procedural rationality, 74–80, 136, 207–209

profiling, in General Data Protection Regulation, 259

proportionality tests, legality principle and, 160

proxies, in algorithmic systems, 62–66, 283

public authorities

accountability and, 218–221

AI Act and, 274–288

definition and scope of, 21

systemic vulnerability and, 221–225

supremacy of coders and, 210–213

public interest, non-arbitrariness principle and, 133–138

public participation, in legality principle, 128–129

public sector

algorithmic regulation in, 69–93

administrative acts, 78–82

algocracy, 82–87, 212

bureaucracy and, 70, 82–87

Public Sector Equality Duty, 193–194

rationality. See also techno-rationality

procedural, 74–80, 136, 207–209

substantive, 74–80, 136–137, 149, 207, 211

rational-legal bureaucracy, 75

Raz, Joseph, 98, 102–103, 171

reasoning

algorithms and, 30

knowledge-driven systems and, 30–32

reinforcement learning, 39

representation. See knowledge representation

Revenue and Customs service (HMRC), U.K., 219

Road to Serfdom (Hayek), 137

Romania, LGBTQ+ rights in, 123–124, 222

rule by law. See also algorithmic rule by law

authoritarianism and, 105, 107

rule of law and, 105–109

formal notions of law, 105–106

rule of law. See also equality before the law; European Union; legal certainty; legality; non-arbitrariness; separation of powers

AI Act and, 16, 265–280, 295

analysis of, 95–150

Aristotle on, 96–97

aspirational quality of, 103–104

authoritarian tendencies and, 9, 12, 17, 105–109, 121–125

constitutional retrogression and, 12–13, 224–226

in British constitutional theory, 97–98

citizenship and, 95

Code of Hammurabi and, 96

common good and, 103, 137–138

conceptualisation of, 95–109

Anglo-American, 97–98

in EU, 109–125

history of, 96

contestation of, 103, 109

Council of Europe and, 106–108, 110–112

crisis

in EU, 121–125, 144

in Israel, 12

critiques of, 102–103

definitions of, 96–109

Dicey on, 97–98

EU legal order and, 104, 112–125, 232–255

formal notions of, 101–105

as thin notion, 101

Fuller on, 99, 144

human rights protections under, 9, 11, 102

in hybrid regimes, 107

illiberal tendencies and, 16, 121–125

constitutional retrogression and, 12–13

judicial review principle, 140–144

for administrative acts, 142

in Belgium, 142–143

challenges to, 142–144

requirements of, 140–142

list-based approaches to, 99–100

naturalism and, 101–102

as normative legal framework, 125–148

positivism and, 101–102

Raz on, 98, 102–103, 171

rule by law and, 105–150

spirit of, 96, 99–101

substantive notions of, 101–105

as thick notion, 101

teleological approach to, 99–101

as theatre, 103

theoretical approaches to, 96–109

Schmitt, Carl, 102–103

Scientific Panel of Independent Experts, 273

self-assessment,

in AI Act, 272, 281–282, 292

separation of powers principle, 144–149

algorithmic regulation and, 199–205

citizen surveillance, 202–205

executive branch and, 200–201

privatisation of legal infrastructure, 201–202

public power, 201–202

media and, 145–146

trias politica doctrine, 145, 201

Shklar, Judith, 98, 103, 158, 170

legalism defined by, 158

societal harm, 7–8, 15, 50, 142, 220, 228

socio-technical infrastructure. See infrastructure

speech acts, 153

spurious correlations, 38, 42–43, 56, 196

errors and, 56

standardisation

in AI Act, 273, 293

organisations, 273

Stalinism, 79, 225

Star, Susan, 54

STIR system, in Poland, 196–197

storage limitation principle, 256–258

Strasbourg case law, 118, 127–129, 139, 141, 144, 191

substantive rationality, 74–80, 136–137, 149, 207, 211

supervised learning, 36

supremacy of coders, 210–213, 222, 294

bureaucratic justice, 211

outsourcing of discretion, 213, 283

surveillance, 66, 202–205

Sweden, algorithmic systems in, 167, 172

symbolic rules, 31–32, 41–43, 46

systemic harm, 198

tax fraud, 196–197, 242–245

knowledge-driven systems and, 32–34

techno-rationality, primacy of, 207–210, 294

interpretation of legal rules and, 208

testing phase, of data-driven systems, 36, 62, 277

TEU. See Treaty on European Union

TFEU. See Treaty on the Functioning of the European Union

Tocqueville, Alexis de, 106–107

traditional AI, 31

training phase, of data-driven systems, 62, 183, 277, 283

transparency. See also opacity

accountability and, 218–219

in administrative acts, 60, 163, 172

under AI Act, 266, 270–277, 284–285, 291–292, 303–305

deep fakes, 276

in data-driven systems, 59–60

judicial review principle and, 191–192

legality principle and, 128–129, 161–164

Treaty on European Union (TEU), 11–14, 114–119, 148, 231–232, 247

Article 2, 11, 127, 138, 232–237, 251–252

Article 7, 223, 233–237, 239–240, 252–255

conferral principle in, 115, 231

national identity in, 148

legality principle and, 127

Treaty on the Functioning of the European Union (TFEU), 239–240, 244–255

AI Act and, 267–268

infringement actions under, 13, 246–255

trias politica doctrine, 145, 201

Trustworthy AI, 266–267, 288

United Kingdom (UK)

British constitutional theory, 97–98

Equality Act 2010, 193

legality principle in, 163–164

Post Office scandal in, 57, 198

Public Sector Equality Duty in, 193–194

Revenue and Customs services in, 219

welfare benefits in, 169

unsupervised learning, 37–38, 42

United States (US)

algorithmic regulation in, 55, 157–158, 174, 183, 185, 200

value-added tax (VAT), 244–245

values

in European Union, 11, 114, 117, 120–125

plurality of, 76

VAT. See value-added tax

Venice Commission, Council of Europe and, 108, 110–111, 119, 127, 133, 141, 144, 160

Weber, Max, 71–78, 208

Weiler, Joseph, 106

welfare benefit allocation, 158, 174

knowledge-driven systems and, 32

in U.K., 169

welfare fraud

data-driven systems and, 37–38

in Indiana, 158

knowledge-driven systems and, 33

in the Netherlands, 187

Dutch Childcare Allowance case, 108, 159–161, 184, 198, 262

welfare state, development of, 136. See also welfare benefit allocation

Yeung, Karen, 180, 212

Book contents

Index

Information

Index

Accessibility standard: Unknown

Book contents

Index

Information

Accessibility standard: Unknown

Save book to Kindle

Save book to Dropbox

Save book to Google Drive