Engaging with the heritage of Nazi persecution in the age of AI: ethical considerations and educational implications

Abstract This paper examines the ethical and educational implications of using AI in digital heritage environments dedicated to the Holocaust and Nazi persecution. Drawing on the Horizon Europe MEMORISE project, it argues that AI can support access to fragmented and multimodal heritage materials through tagging, linking, translation, recommendation, and personalization, but only if these functions remain embedded in a human-centered framework grounded in dignity, authenticity, transparency, provenance, and historical accuracy. The paper highlights the risks of algorithmic bias, decontextualization, hallucination, artificial empathy, and over-personalized “filter bubbles,” especially in relation to sensitive historical sources and marginalized victim perspectives. In response, it proposes an approach based on human-in-the-loop oversight, provenance by design, AI literacy, and experiential knowledge, showing how AI can enhance critical engagement with difficult histories without replacing human interpretation, educator responsibility, or the integrity of archival sources.