Publication
This report presents key interdisciplinary insights from IRGC’s expert workshop on the governance of decision-making algorithms, with particular focus on automated decisions based on learning algorithms (DMLAs). It highlights, among others, the need to improve the accuracy of DMLAs; embed good standards, principles or governance rules into their very design and functioning; address algorithmic bias with unfair social consequences; better differentiate if and when humans are in control, or unable to take control back; govern distinct risks arising by application or domain (e.g. DMLAs in healthcare/automated driving/predictive policy-making/insurance, etc.); better state key benchmarks against which DMLAs’ performance must be calibrated; better define accountability, responsibility and liabilities; and emphasize the centrality of digital trust.