With the help of artificial intelligence-powered services and tools the heritage sector is working towards the next level of access to and (re)use of digitized collections. In recent years libraries, archives and museums have started to apply machine learning and advanced knowledge bases to contextually enrich digitized objects, audio-visual content and texts and to make these retrievable in novel ways. In doing so institutions aim to increase the impact of their collections among a growing and diversifying audience. This special session will include papers that reflect upon, discuss and present the technical and societal challenges (e.g. labour to produce labeled datasets, heterogeneity of data, bias in training sets) digital heritage professionals and researchers are facing when trying to capitalise on the transformative power of artificial intelligence in the context of digital archive, image, and audio/visual collections.