Towards AI ethics-led sustainability frameworks and toolkits: Review and research agenda

Artificial intelligence (AI) is instrumental in building human skills, accessing knowledge, creating businesses, addressing societal concerns–including environmental issues–and much more. However, unfair, inequitable, and biased data usage for AI deployments does exist and raises ethical and sustainability debates and concerns. AI deployment frameworks are majorly developed by standard societies/groups, technology organisations, analyst groups and federal/government agencies. The paper explores the central themes of AI ethics and sustainability frameworks in declarative standards and statements published by various institutions. The paper offers a thematic analysis of the literature on AI ethics-led sustainability frameworks using MAXQDA software and identifies common principles. We show that there are an established 28 AI ethics-led sustainability frameworks that agencies and groups have disseminated. As well, 6 practical AI ethics toolkits/products are evaluated to translate common AI ethics-led sustainability framework recommendations to deploy AI ethics-led sustainability toolkits programmatically. The research findings validate that beneficence, non-maleficence, justice, explainability, autonomy, privacy, and biasedness need severe attention and postulating algorithmic trust based on AI ethics-led sustainability frameworks. The paper contributes to the unique AI ethics-led sustainability body of knowledge to become a helpful resource for both praxis and researchers.