Responsible AI Platform
For Executive Agencies

AI Act for UWV, SVB & Executive Agencies

From unemployment benefits to student finance — how does your organization comply with the AI Act?

Specific guidance for executive agencies with large-scale, impactful AI systems.

Directly applicable for public bodies and executive agencies

This Applies to Your Organization

UWV

Unemployment benefits, disability, WIA assessments

SVB

State pension, child benefits, personal budget

Tax Authority

Benefits, fraud detection, risk selection

DUO

Student finance, diploma registration, recovery

RVO

Subsidies, permits, business regulations

CJIB

Collection, fine processing, payment arrangements

High-Risk AI at Executive Agencies

Many systems at executive agencies fall under Annex III of the AI Act

Benefit Assessments

Automated assessment of applications for unemployment, disability or benefits

Direct influence on citizens' livelihoods

Fraud Risk Profiling

Algorithms that determine which files receive extra scrutiny

Risk of discrimination and unequal treatment

Recovery Decisions

Automated calculation and decision-making on recoveries

Financial consequences for vulnerable groups

Triage and Prioritization

AI-driven work distribution and urgency determination

Determines access to services

Lessons from the Past

Executive agencies have learned important lessons

Tax Authority

Benefits Scandal

How risk profiling and automated decisions led to great injustice.

Lesson:

Human oversight and explainability are essential

Various organizations

SyRI Ruling

System Risk Indication was banned by the court.

Lesson:

Privacy and proportionality must be ensured

UWV

Algorithmic Decision-Making

Research showed that not all systems were traceable.

Lesson:

Transparency and documentation are mandatory

Specific Obligations

What the AI Act requires from executive agencies

Human Oversight

Decisions affecting citizens must always involve a human. "Meaningful human oversight" is mandatory.

Risk Management System

Continuous monitoring and evaluation of AI systems for risks and impact.

Technical Documentation

Extensive documentation on operation, training data and decision-making logic.

Conformity Assessment

High-risk systems must undergo a conformity assessment.

Need Help?

We help executive agencies with practical AI Act implementation.

30-minute orientation call