The Amsterdam Paper: Recommendations for the technical finalisation of the regulation of GPAI in the AI Act

    The vivid discussions around possible approaches to govern General Purpose AI (GPAI) over the past months in and outside Europe have demonstrated the extent of complexity of this task, the level of uncertainty and lack of ready-made and tested solutions, but also the value of joining forces and bringing together the expertise of regulators, civil society and researchers from various disciplines.

    In the paper below, Natali Helberger, Laurens Naudts, and Stanislaw Piasecki aim to summarise the main insights from a workshop that brought together a group of academics and civil society representatives in Amsterdam. The objective of the workshop was to share and learn from each other’s insights and expertise and jointly brainstorm on the problem of how to govern GPAI and what are possible routes for the European AI Act (AIA). The workshop was held in July 2023 and co-organised by AlgorithmWatch and the AI, Media & Democracy Lab of the University of Amsterdam.

    We publish the report in the hope that it can be a useful source of information and critical reflection for finalising and operationalising the European governance approach to GPAI. Given their specific focus on GPAI, the recommendations in this report are not comprehensive but should be regarded as complementary to other recommendations on the AIA.

    Read and download the full paper below.