This tool, developed with AlgorithmWatch and AlgorithmWatch Switzerland, provides a practical framework for evaluating automated decision-making systems (ADMS) in public administration. Now in its fourth version (updated January 2025), it consists of two parts: a triage checklist to determine which systems require a transparency report, and an adaptive accountability and transparency reporting section. The tool is agnostic to specific auditing methodologies, making it applicable across different contexts and levels of public authority.
The tool can be accessed directly at algorithmwatch.tech/impact-assessment-en/.
Related work: together with ECNL and the Danish Institute for Human Rights, I contributed to a Guide to Fundamental Rights Impact Assessments (FRIA) developed in the context of the EU AI Act — a complementary framework for rights-based impact assessment.
