Advances in computing hardware and software along with the ever-increasing availability of huge amounts of data are enabling the deployment of machine learning and statistical model algorithms to solve a range of complex problems and to underpin more and more real-life applications. As our societies become more reliant on technology-enabled and computer devices in general, how can we ensure that this does not increase our power consumption? In this workshop, we will discuss the energy costs associated with running more and more computer-intensive, data-greedy models, possible model optimisation approaches to make best-use on new computer infrastructures, and the development of low-power computer chips designed specifically with AI applications in mind.