Signaling transparency in the era of artificial intelligence

Purpose This study provides researchers and business practitioners with a comprehensive understanding of artificial intelligence (AI) transparency in the business discipline, enabling them to navigate the evolving digital landscape, where AI transparency is an escalating concern, by identifying the conceptual foundations in the most influential studies. Design/methodology/approach This study uses bibliometric analysis techniques, including performance and co-citation analyses. These analyses are grounded in data extracted from the Social Sciences Citation Index within the Web of Science, comprising 108 primary articles and 7,459 secondary (cited) documents. Findings AI transparency research is rising with a greater focus on end-users. Six clusters of cited publications serve as the bedrock of AI transparency in the business discipline: trust, AI explanation, bias and power, undesirable usage, user acceptance/aversion and user heuristics. Analyzing these clusters revealed a framework for signaling AI transparency that can be extended to future research and business strategies. Originality/value This study addresses the following research gaps. First, the nature of AI transparency and its knowledge base remain elusive. Second, AI transparency in the business discipline is underexplored compared to information sciences and law. Third, there is ambiguity surrounding the implementation strategies for AI transparency, with companies often resorting to simplistic methods such as updating terms and conditions. Fourth, there is a lack of clear future research directions specifically for AI transparency, as opposed to the broader context of AI ethics.