Activity: Talks and presentations › Conference presentations
We’re in the middle of a ‘data revolution’: large, detailed datasets and complex algorithms are being deployed to make predictions from which investments lead to economic growth to who is likely to commit a further crime, and to automate processes from the detection of malignant moles to the elimination of bad software code. So how do we expand our ability to question the quality of evidence to meet this - as the public, research bodies, commissioners or decision makers? How do we determine the reliability of data science or AI? What kind of testing should be expected, and what are our reliability thresholds? Swamped by concerns about privacy and disruption, issues of quality and rigor in data science have largely been neglected. These are essential questions. We cannot meaningfully assess applications without them, or ask whether we are maximising data science benefits for research and society.