The Role of Data in AI: Report for the Data Governance Working Group of the Global Partnership of

This is the final report of the project The Role of Data in AI, which was commissioned by the Data Governance Working Group (WG) of the Global Partnership of AI (GPAI). The overarching aim of the project was to highlight and describe the role of data in AI development processes and identify key challenges related to data data quality, accessibility and availability. We also describe the impact these challenges have on AI development, at societal and individual levels. The Role of Data in AI project ran between 17th September - 7th December 2020 and was led by the Digital Curation Centre, with project partners Trilateral Research and School of Informatics, The University of Edinburgh. The report is based on a review of literature and consultation with expert members of GPAI and the Data Governance WG through a series of three workshops and weekly meetings. The first three sections of the report describe the role of data in AI development as well as key types of data that are used and their characteristics. It highlights the importance of having vast amounts of good quality data for AI development for best results and how data limitations can lead to poor results, which can have negative impacts on society and individual rights. Section 5 goes into more depth and examines data-related issues emerging from the collection, process and use of data in AI and offers a wide mapping of important issues to inform the further developments of AI creation. It provides a brief examination of the impact of access to datasets and use of different types of data for the creation of AI.