| Popis: |
As of 2021, there were more than 170 guidelines on AI ethics and responsible, trustworthy AI in circulation according to the AI Ethics Guidelines Global Inventory maintained by AlgorithmWatch, an organisation which tracks the effects of increased digitalisation on everyday lives. However, from the perspective of day-to-day work, for those engaged in designing, developing, and maintaining AI systems identifying relevant guidelines and translating them into practice presents a challenge.The aim of this paper is to help anyone engaged in building a responsible AI system by identifying an indicative long-list of potential stakeholders. This list of impacted stakeholders is intended to enable such AI system builders to decide which guidelines are most suited to their practice. The paper draws on a literature review of articles short-listed based on searches conducted in the ACM Digital Library and Google Scholar. The findings are based on content analysis of the short-listed literature guided by probes which draw on the ISO 26000:2010 Guidance on social responsibility.\ud \ud The paper identifies three levels of potentially relevant stakeholders when responsible AI systems are considered: individual stakeholders (including users, developers, and researchers), organisational stakeholders, and national / international stakeholders engaged in making laws, rules, and regulations. The main intended audience for this paper is software, requirements, and product engineers engaged in building AI systems. In addition, business executives, policy makers, legal/regulatory experts, AI researchers, public, private, and third sector organisations developing responsible AI guidelines, and anyone interested in seeing functional responsible AI systems are the other intended audience for this paper. |
