Many social and scientific domains give rise to data with multi-way relationships that can naturally be represented by tensors, or multi-dimensional arrays. Decomposing - or factoring - tensors can reveal latent properties that are otherwise difficult to see. Tensor decomposition is currently widely used in applications such as signal processing, recommender systems, and topic modeling. In particular, topic modeling, where abstract ‘topics’ can be identified in a collection of documents, can be used to determine the latest areas of interest and of importance in AI from online forums and websites, such as Reddit and Twitter. This may aid in identifying the appropriate topics for leading the discussing in advancing AI, for both research and for social impact.