The predominant method to analyse the morphology of galaxies is to use human visual inspection. The Galaxy Zoo project has successfully scaled this process to incorporate crowd sourced inspection and has released detailed classifications of hundreds of thousands of galaxies. However, the next generation of telescopes are now being designed and built. An optical telescope called the Large Synoptic Sky Telescope, for example, is due to come online in 2019 and will image half the sky every few days to catalogue ~40 billion objects. This is beyond the capability of even the significant crowd sourced resources of Galaxy Zoo. Therefore, astronomers are now investigating automated solutions using machine learning. In this talk I outline the key issues that need to be resolved in order to automate galaxy morphological analysis in large surveys. I describe existing automated systems that analyse galaxies, and I describe the results of my work using unsupervised machine learning approaches to characterize galaxies.