Guide to algorithmic impact assessment on fundamental rights

Practical guide for teams that need to assess the impact of an artificial intelligence system on people’s rights before putting it into production.

The document walks through the whole process: defining the system and its context of use, identifying the affected groups, analysing the rights that may be compromised, assessing the likelihood and severity of the impact, and defining mitigation measures with assigned owners.

Contents

Conceptual framework and regulatory fit. Assessment template in editable format. Criteria for deciding when an assessment must be repeated. Annotated examples in three sectors: employment, social services and education.

Sample document for repository layout.