Algorithmic Memory: Towards Reflexive Authenticity in Cultural Heritage

Abstract

Artificial intelligence is reshaping cultural heritage not only as a technological instrument of preservation but as a philosophical framework that transforms how the past is remembered, interpreted, and curated. Through text reconstruction, corpus analysis, and language restoration, AI expands the analytical and participatory capacities of digital heritage while simultaneously challenging established epistemologies of authenticity and authority (Jones et al., 2021; Floridi, 2019). This article proposes the Reflexive Authenticity Framework, an ethical and methodological model that redefines authenticity as transparency and curatorial authority as participatory rather than hierarchical. Grounded in algorithmic literacy, inclusive governance, and ethical auditing (UNESCO, 2021), the framework promotes epistemic integrity in AI-mediated heritage environments. Case studies including Europeana, the Perseus Digital Library, and Mukurtu CMS illustrate forms of human–machine co-authorship that generate hybrid cultural memory. At the same time, the article identifies the risk of epistemic drift, whereby unverified synthetic narratives circulate faster than critical validation, underscoring the need for reflexive verification practices in digital historiography.