The ethical dimensions of Artificial Intelligence (AI) constitute a salient topic in information systems (IS) research and beyond. There is an increasing number of journal and conference articles on how AI should be designed and used. For this, IS research offers and curates knowledge not only on the ethical dimensions of information technologies but also on their acceptance and impact. However, the current discourse on the ethical dimensions of AI is highly unstructured and seeks clarity. As conventional systematic literature research has been criticized for lacking in performance, we applied an adapted discourse approach to identify the most relevant articles within the debate. As the fundamental manuscripts within the discourse were not obvious, we used a weighted citation-based technique to identify fundamental manuscripts and their relationships within the field of AI ethics across disciplines. Starting from an initial sample of 175 papers, we extracted and further analyzed 12 fundamental manuscripts and their citations. Although we found many similarities between traditionally curated ethical principles and the identified ethical dimensions of AI, no IS paper could be classified as fundamental to the discourse. Therefore, we derived our own ethical dimensions on AI and provided guidance for future IS research.