Why Value Judgements Should Not Be Automated

AI technologies are already being used for a number of purposes in public services, including to automate (parts of) decision processes and to make recommendations and predictions in support of human decisions. Increasing application of AI in public services therefore has the potential to impact several of the Seven Principles of Public Life, presenting new challenges for public servants in upholding those values. We believe AI is particularly likely to impact the principles of Objectivity, Openness, Accountability and Leadership. Algorithmic bias has the potential to threaten the objectivity of public sector decisions, while several forms of opacity in AI systems raise challenges for openness in public services; and, in turn, this impacts the ability of public servants to be accountable and exercise proper leadership.Leverhulme Trust Centre Grant (Leverhulme Centre for the Future of Intelligence