Recent interest, investment in and concern with artificial intelligence, as a strategic area of research and innovation, is connecting disciplines across the sciences and humanities in potentially new ways. Much is expected, in terms of the contribution to knowledge, economy and society, from new applications of computational methods such as machine learning to social and cultural topics. This has led some thinkers to express concern that the capacity to know society and culture is being shifted to disciplines with little experience and training in the distinctive challenges that have been associated with these projects. However, scholars and scientists have also argued that recent developments in artificial intelligence make possible fundamental transformations of the relations between the sciences, social sciences and humanities, enabling new ways of combining approaches, ontologies and methodologies, which have historically been considered incommensurable (Rawhan et al, 2019; Castelle, forthcoming).1 In this workshop, we would like to examine these challenges and opportunities for interdisciplinary exchange in AI research and innovation with the aid of a diverse and experimental set of evaluative methods: scientometrics, evaluative enquiry, data visualisation and playful methods.