Table A1 Dimensions and indicators used...

Table A1

Dimensions and indicators used for policy coding

Dimension	Indicator	Operational definition
D1 – Social Risk Allocation	State responsibility	Assigns primary responsibility to government/public authorities for mitigating AI risks
	Employer obligations	Places compliance or risk-management duties on firms, deployers, or organisations
	Individual responsibility	Frames users or workers as bearing primary responsibility for managing risks
	Welfare safeguards	Provides social protection, safety nets, or compensatory support measures
	Risk redistribution	Redistributes AI-related risks collectively rather than concentrating them on individuals
D2 – Regulatory Authority and Enforcement	Binding law	Establishes legally binding rules or statutory obligations
	Enforcement body	Designates authorities with oversight or supervisory powers
	Sanctions/penalties	Provides fines, penalties, or corrective measures for non-compliance
	Audit/assessment	Requires audits, testing, conformity assessment, or evaluation procedures
	Monitoring/KPIs	Mandates ongoing reporting, monitoring, or performance indicators
	Risk classification	Categorises AI systems by risk level to determine regulatory treatment
D3 – Labour and Welfare Integration	Labour law reference	Links AI governance to employment or Labour legislation
	Worker protection	Protects workers from displacement, discrimination, or algorithmic harm
	Reskilling/training	Supports education, retraining, or workforce transition measures
	Welfare system linkage	Connects AI governance to social security or welfare institutions
	Collective rights	Recognises trade unions, social dialogue, or collective bargaining rights
D4 – Ethics Institutionalisation	Ethics principles	Articulates normative values (e.g. fairness, transparency, human oversight)
	Procedural ethics	Embeds ethics in organisational processes or governance structures
	Mandatory impact assessment	Requires formal AI or algorithmic impact assessments
	Compliance/audit	Establishes mechanisms to verify adherence to standards or rules
	Legal enforceability	Makes ethical or governance requirements legally binding
	Contestability/remedy	Provides rights or procedures to challenge automated decisions

Dimension	Indicator	Operational definition
D1 – Social Risk Allocation	State responsibility	Assigns primary responsibility to government/public authorities for mitigating AI risks
	Employer obligations	Places compliance or risk-management duties on firms, deployers, or organisations
	Individual responsibility	Frames users or workers as bearing primary responsibility for managing risks
	Welfare safeguards	Provides social protection, safety nets, or compensatory support measures
	Risk redistribution	Redistributes AI-related risks collectively rather than concentrating them on individuals
D2 – Regulatory Authority and Enforcement	Binding law	Establishes legally binding rules or statutory obligations
	Enforcement body	Designates authorities with oversight or supervisory powers
	Sanctions/penalties	Provides fines, penalties, or corrective measures for non-compliance
	Audit/assessment	Requires audits, testing, conformity assessment, or evaluation procedures
	Monitoring/KPIs	Mandates ongoing reporting, monitoring, or performance indicators
	Risk classification	Categorises AI systems by risk level to determine regulatory treatment
D3 – Labour and Welfare Integration	Labour law reference	Links AI governance to employment or Labour legislation
	Worker protection	Protects workers from displacement, discrimination, or algorithmic harm
	Reskilling/training	Supports education, retraining, or workforce transition measures
	Welfare system linkage	Connects AI governance to social security or welfare institutions
	Collective rights	Recognises trade unions, social dialogue, or collective bargaining rights
D4 – Ethics Institutionalisation	Ethics principles	Articulates normative values (e.g. fairness, transparency, human oversight)
	Procedural ethics	Embeds ethics in organisational processes or governance structures
	Mandatory impact assessment	Requires formal AI or algorithmic impact assessments
	Compliance/audit	Establishes mechanisms to verify adherence to standards or rules
	Legal enforceability	Makes ethical or governance requirements legally binding
	Contestability/remedy	Provides rights or procedures to challenge automated decisions

[ViewLarge]

Sharing Unavailable