An emerging topic in artificial intelligence (AI) projects is project teams’ responsibilities for developing ethical AI systems. Consequently, there are many ethical guidelines and principles that project teams should take into consideration. For example, Jobin, Ienca, and Vayena performed a content analysis of 84 AI ethical guidelines. 84. Ryan and Stahl translated those guidelines into 11 ethical principles that users and developers should use to take moral responsibility for the AI systems they develop.
We expanded the work from Ryan and Stahl to translate the ethical principles into project success factors that include project deliverable and actions. We attempted to consider success from the perspective of internal stakeholders that would finance and benefit from the systems and other stakeholders that could be impacted by the systems. You can find the five categories, 17 groups, and 84 success factors in our article: Artificial Intelligence Project Success Factors—Beyond the Ethical Principles. In this blog post, we would like to expand on an abbreviated version of the heatmap from the article.
Figure 1 is a heatmap that shows the intersection between project success and AI ethical principles. The project success categories and groups are along the Y-axis, AI ethical principles along the X-axis, and the cells include a count for the number of success factors at the intersection. A detailed explanation of each success category, group, and factor is included in the article, which you can download from the Springer site.