How can equality bodies determine whether an algorithmic or artificial-intelligence (AI) system has contributed to discrimination – and what should be the steps taken to find redress? How can public administrations ensure that the development and use of AI go hand-in-hand with effective protection of equality and non-discrimination? The newly published European assessment methodology on algorithmic discrimination translates standards and legal and technical expertise into a practical, step-by-step approach for practitioners. It aims to provide answers.

The new publication, authored by Louise Hooper, Ylja Remmits and Ioanna Papageorgiou, is a tool to support equality bodies and other actors in identifying the algorithmic systems involved in a case, gathering the information necessary to identify potential discrimination, and assessing it from legal and technical perspectives. It will also help identify possible remedies when discrimination is found.

Cooperating across the continent

The assessment technology was presented today at the closing event of the Technical Support Instrument project “Upholding equality and non-discrimination by Equality bodies regarding the use of AI in public administrations”, co-funded by the European Union and implemented by the Council of Europe, in cooperation with the European Commission.

Over two years, equality bodies from three European countries (Unia from Belgium, the Non-Discrimination Ombudsman from Finland, and the Commission for citizenship and gender equality from Portugal) have worked together with the project experts. The objective has been to develop tools and practices to support public administrations across Europe in adjusting to the rapidly evolving European regulatory landscape. The assessment methodology published today is the project’s newest publication; it is available in English, Dutch, French, Swedish, Finnish and Portuguese.

 Find out more about the Council of Europe’s world-leading work on AI