MIT researchers show how to detect and address AI bias without loss in accuracy

MIT CSAIL researchers have developed methods to reduce bias in predictive AI without the loss of accuracy. Full results will be shared next month at NIPS.