D5.8 Artificial Intelligence Impact Assessment - A Systematic Review

Artificial intelligence (AI) is expected to produced impacts that are highly beneficial, but it also raises concerns about undesirable ethical and social consequences. There is an array of activities that aim to address these undesirable consequences, ranging from proposals for regulation such as the EU AI Act to ethics guidelines to design methodologies, professional guidance or standardisation. One option that is increasingly explored is to develop impact assessments specifically geared for the needs of AI. A number of such AI impact assessments (AI-IAs) have already been proposed. This document undertakes a systematic review of these AI-IAs with the aim of identifying whether there are common themes and approaches. This research is important to establish a baseline for AI-IAs that can help organisations identify AI-IAs that are most relevant to their needs and that can serve as a measure for legislators and regulators to determine the role that AI-IAs can play in the governance of the broader AI ecosystem.