A learning approach to responsible AI – the ELSA-lab (ethical legal and societal aspects) methodology

Bram Klievink (Leiden University)

Roel Dobbe (Delft University)

Jord Goudsmit (Considerati)

Anne Fleur van Veenstra (TNO)

Marianne Schoenmaker (Considerati)

ELSA stands for Ethical, Legal and Societal Aspects. This great collection of disciplines is a good foundation for developing labs on. The position paper ‘ELSA Labs for Human Centric Innovation in AI’ explains the background and the value of the ELSA concept for the desired joint approach, in which the general public also participate actively. It discusses experiences with the ELSA concept over the past twenty years as well as the practical risks and challenges if a rapid learning curve is to be achieved for deploying the ELSA Labs.

The aim is to ensure that companies, governmental authorities, centres of expertise, civil society organisations and the general public develop responsible applications of AI jointly. This involves solutions for both social and business problems, focusing on honesty, fairness, security and (above all) trustworthiness. The approach addresses human values as well as public values.

All the stakeholders are involved in developing human centric AI with clear ethical and legal frameworks and helpful regulations within the European context so that the socioeconomic impact of AI becomes manageable and people will be confident about how it works. The Human Centric AI working group is aiming to create thirty of these ELSA Labs in the period through to 2027. In January 2022, after review by an independent evaluation committee from the Netherlands Organisation for Scientific Research (NWO), the first five labs from the Dutch National Research Agenda call ‘Human-Centric AI for an inclusive society: Towards an ecosystem of trust’ were awarded funding.