LIST OF ABBREVIATIONS 7

LIST OF BOXES 11

LIST OF FIGURES 11

LIST OF TABLES 11

EXECUTIVE SUMMARY 13

1. INTRODUCTION 16

  1.1. Scope and structure 16

  1.2. Method 17

2. COUNTRY PROFILES 18

  2.1. Türkiye 18

    2.1.1. Role at COP31 19

    2.1.2. Policy and emissions context 20

    2.1.3. Adaptation and resilience 20

    2.1.4. Climate finance 21

    2.1.5. Negotiating position 21

    2.1.6. Progress on selected voluntary commitments 22

  2.2. Australia 24

    2.2.1. Role at COP31 25

    2.2.2. Policy and emissions context 26

    2.2.3. Adaptation and resilience 26

    2.2.4. Climate finance 27

    2.2.5. Negotiating position 27

    2.2.6. Progress on selected voluntary commitments 29

  2.3. Bangladesh 30

    2.3.1. Policy and emissions context 30

    2.3.2. Adaptation and resilience 31

    2.3.3. Climate finance 32

    2.3.4. Negotiating position 33

  2.4. Brazil 35

    2.4.1. Policy and emissions context 35

    2.4.2. Adaptation and resilience 36

    2.4.3. Climate finance 37

    2.4.4. Negotiating position 37

  2.5. China 39

    2.5.1. Policy and emissions context 39

    2.5.2. Adaptation and resilience 40

    2.5.3. Climate finance 41

    2.5.4. Negotiating position 42

  2.6. Colombia 43

    2.6.1. Policy and emissions context 43

    2.6.2. Adaptation and resilience 44

    2.6.3. Climate finance 44

    2.6.4. Negotiating position 45

  2.7. India 46

    2.7.1. Policy and emissions context 46

    2.7.2. Adaptation and resilience 47

    2.7.3. Climate finance 48

    2.7.4. Negotiating position 48

  2.8. Indonesia 50

    2.8.1. Policy and emissions context 50

    2.8.2. Adaptation and resilience 51

    2.8.3. Climate finance 52

    2.8.4. Negotiating position 53

  2.9. Japan 54

    2.9.1. Policy and emissions context 54

    2.9.2. Adaptation and resilience 55

    2.9.3. Climate finance 56

    2.9.4. Negotiating position 56

  2.10. Kenya 58

    2.10.1. Policy and emissions context 58

    2.10.2. Adaptation and resilience 59

    2.10.3. Climate finance 60

    2.10.4. Negotiating position 61

  2.11. Maldives 62

    2.11.1. Policy and emissions context 62

    2.11.2. Adaptation and resilience 63

    2.11.3. Climate finance 64

    2.11.4. Negotiating position 65

  2.12. Mexico 66

    2.12.1. Policy and emissions context 66

    2.12.2. Adaptation and resilience 67

    2.12.3. Climate finance 68

    2.12.4. Negotiating position 69

  2.13. Nigeria 70

    2.13.1. Policy and emissions context 70

    2.13.2. Adaptation and resilience 71

    2.13.3. Climate finance 72

    2.13.4. Negotiating position 73

  2.14. Philippines 74

    2.14.1. Policy and emissions context 74

    2.14.2. Adaptation and resilience 75

    2.14.3. Climate finance 75

    2.14.4. Negotiating position 76

  2.15. Saudi Arabia 78

    2.15.1. Policy and emissions context 78

    2.15.2. Adaptation and resilience 79

    2.15.3. Climate finance 80

    2.15.4. Negotiating position 81

  2.16. South Africa 82

    2.16.1. Policy and emissions context 82

    2.16.2. Adaptation and resilience 83

    2.16.3. Climate finance 84

    2.16.4. Negotiating position 85

  2.17. Uganda 86

    2.17.1. Policy and emissions context 86

    2.17.2. Adaptation and resilience 88

    2.17.3. Climate finance 89

    2.17.4. Negotiating position 90

  2.18. United Arab Emirates 91

    2.18.1. Policy and emissions context 91

    2.18.2. Adaptation and resilience 92

    2.18.3. Climate finance 93

    2.18.4. Negotiating position 94

  2.19. United Kingdom 95

    2.19.1. Policy and emissions context 95

    2.19.2. Adaptation and resilience 96

    2.19.3. Climate finance 97

    2.19.4. Negotiating position 98

  2.20. United States 101

    2.20.1. Policy and emissions context 101

    2.20.2. Adaptation and resilience 102

    2.20.3. Climate finance 104

    2.20.4. Negotiating position 105

3. PARTY GROUPINGS 107

  3.1. Group of 77 and China (G77+China) 108

  3.2. African Group of Negotiators (AGN) 109

  3.3. Independent Association of Latin America and the Caribbean (AILAC) 109

  3.4. Alliance of Small Island States (AOSIS) 110

  3.5. Arab Group 111

  3.6. Environmental Integrity Group (EIG) 112

  3.7. Least Developed Countries Group (LDC Group) 112

  3.8. Like-Minded Developing Countries (LMDC) 113

  3.9. Umbrella Group 114

4. CIVIL SOCIETY CONSTITUENCIES AND REGIONAL ORGANISATIONS 115

  4.1. Indigenous Peoples Organisations (IPO) 115

    4.1.1. Institutional role and representation 115

    4.1.2. Positions on key negotiation issues 116

  4.2. Children and Youth Constituency (YOUNGO) 117

    4.2.1. Institutional role and representation 117

    4.2.2. Positions on key negotiation issues 117

  4.3. Women and Gender Constituency (WGC) 118

    4.3.1. Institutional role and representation 118

    4.3.2. Positions on key negotiation issues 119

  4.4. Climate Action Network (CAN) 120

    4.4.1. Institutional role and representation 120

    4.4.2. Positions on key negotiation issues 120

  4.5. Loss and Damage Collaboration (L&DC) 121

    4.5.1. Institutional role and representation 121

    4.5.2. Positions on key negotiation issues 121

  4.6. Association of Southeast Asian Nations (ASEAN) 122

    4.6.1. Institutional role and representation 122

    4.6.2. Positions on key negotiation issues 122

REFERENCES 124

LIST OF ABBREVIATIONS 6

LIST OF BOXES 9

LIST OF TABLES 9

EXECUTIVE SUMMARY 10

1. INTRODUCTION 13

  1.1. Research questions 13

  1.2. Methodology 14

  1.3. Structure of the study 14

  1.4. Scope 14

2. AI AND GENDER BIAS: CONCEPTS, DEFINITIONS AND SOURCES 16

  2.1. The concept of artificial intelligence 16

  2.2. The definition of AI at the EU level 17

  2.3. The concept of gender bias in AI 18

  2.4. The sources of gender bias in AI 20

    2.4.1. Pre-existing bias 21

    2.4.2. Technical bias 21

    2.4.3. Emergent bias 23

    2.4.4. Concluding considerations 24

3. HOW GENDER BIAS IN AI AFFECTS WOMEN: EVIDENCE FROM FOUR DOMAINS 25

  3.1. Employment and the labour market 26

    3.1.1. Exposure to AI and equal access 26

    3.1.2. Gender bias from recruitment to workplace 27

    3.1.3. Intersecting inequalities in the automated workplace 28

  3.2. Public and private services 28

    3.2.1. Welfare and social security 28

    3.2.2. Credit scoring, lending, and insurance 29

  3.3. Healthcare 30

    3.3.1. Gender bias in clinical and administrative uses of AI 30

    3.3.2. The case of femtech 31

  3.4. Social media 32

    3.4.1. Recommender systems 32

    3.4.2. Content moderation 33

    3.4.3. Advertising systems 33

  3.5. Other domains 34

4. TECHNOLOGY-FACILITATED GENDER-BASED VIOLENCE: A PERVASIVE AND EVOLVING PHENOMENON 35

  4.1. Terminology 35

  4.2. Manifestations 37

    4.2.1. Image-based sexual abuse 37

    4.2.2. Sexual digital forgeries 38

    4.2.3. Sextortion 39

    4.2.4. Cyberharassment 40

    4.2.5. Cyberstalking 40

    4.2.6. Gender-based hate speech 41

    4.2.7. Meta-rape 42

    4.2.8. Chatbot-simulated and sex robot-mediated abuse 42

  4.3. Nature and impact 42

  4.4. Scale and data collection 44

5. LEGAL FRAMEWORK 46

  5.1. The anti-discrimination framework 47

    5.1.1. The scope of EU gender equality and anti-discrimination law 47

    5.1.2. Prevailing forms of algorithmic discrimination 48

  5.2. The European Union’s digital acquis 51

    5.2.1. The Artificial Intelligence Act 51

    5.2.2. The General Data Protection Regulation 55

    5.2.3. The Digital Services Act 56

    5.2.4. The Directive on combatting violence against women and domestic violence 56

    5.2.5. Recent EU legislative and regulatory developments 57

    5.2.6. The Council of Europe’s Framework 59

  5.3. Implementation challenges 59

6. MAPPING OF TECHNICAL, ORGANISATIONAL AND GOVERNANCE MEASURES TO PREVENT AND MITIGATE GENDER BIAS 61

  6.1. AI Act requirements 62

  6.2. Socio-technical measures to prevent and mitigate gender bias 63

    6.2.1. Bias mitigation measures 63

    6.2.2. Fairness metrics 68

  6.3. Organisational and governance measures to prevent and mitigate gender bias 70

    6.3.1. Fundamental rights and equality impact assessments, participation and prevention 70

    6.3.2. The use of AI to advance gender equality 72

  6.4. Selection of best practices at European and national level 73

7. REDRESS, OVERSIGHT AND ACCOUNTABILITY MECHANISMS AND THEIR EFFECTIVENESS 75

  7.1. Barriers to effective redress for AI-related harm 76

    7.1.1. Barriers relating to the identification and reporting of harm 76

    7.1.2. Barriers to redress arising from the intersectional, systemic, collective and societal nature of algorithmic harm 77

    7.1.3. Resource and expertise barriers 78

    7.1.4. Institutional and procedural barriers, including transnational dimension 79

  7.2. Emerging rights and mechanisms addressing technology-specific barriers to redress 80

    7.2.1. Information and transparency rights 80

    7.2.2. Complaint and individual redress mechanisms 82

    7.2.3. Collective redress mechanisms 83

    7.2.4. Oversight and enforcement through and in cooperation with specialised actors 84

    7.2.5. Accountability through self-assessment and risk management 85

    7.2.6. Liability, causation and evidentiary burdens 85

8. RECOMMENDATIONS 88

REFERENCES 91

ANNEX 1: METHODOLOGICAL APPROACH 122

ANNEX 2: CONSULTATION QUESTIONNAIRE ON GENDER BIAS AND ARTIFICIAL INTELLIGENCE 124

ANNEX 3: SIMPLIFIED VERSION OF THE QUESTIONNAIRE 126