An Ethical Reflection Aid for Responsible AI in Computational Archival Science

AI implementations continue to grow in cultural heritage settings and have deep connections to how computational archival science is conducted. This paper reviews how AI is being implemented in computational archival science. It then provides an overview of ethical issues and considerations when implementing AI for computational archival science. Our research team has developed an evidence-based ethical reflection aid to guide library and archives practitioners through responsible implementations of AI. When using the ethical reflection aid practitioners consider a specific AI implementation scenario, consider how the different values held by different stakeholders may align or come into conflict, and outline potential actions for responsible AI. We present an example of using the ethical reflection aid to examine a scenario related to computational archival science. Ultimately, this paper suggests that careful examination of stakeholder values can support a more responsible computational archival science practice.