20 Pages Posted: 28 Oct 2016
Date Written: October 27, 2016
As machine learning (ML) becomes increasingly prevalent, concerns are mounting over its use. This discussion paper explores notions of responsibility with regard to ML, focusing on transparency and control. We recognise that such concerns extend beyond the ML technology itself, to the workflows and processes in which the ML operates, i.e. its potential impact. As such, it is important to consider not only the nature of machine learning techniques, but also the data involved and its fit within a broader process. Each of these aspects relate to responsibility, as they represent points for choice and intervention.
Keywords: machine learning, law, responsibility, data, workflow, processes, human in-the-loop, audit, compliance, computer science, work
Suggested Citation: Suggested Citation
Singh, Jatinder and Walden, Ian and Crowcroft, Jon and Bacon, Jean, Responsibility & Machine Learning: Part of a Process (October 27, 2016). Available at SSRN: https://ssrn.com/abstract=2860048