Stuart Mills, Samuël Costa & Cass R. Sunstein, The Opportunities and Costs of AI in Behavioural Science (June 26, 2023).
Abstract: This article discusses the opportunities and costs of AI in behavioural science. We argue that because of pattern detection capabilities, modern AI will be able to identify (1) new biases in human behaviour and (2) known biases in novel situations. AI will also allow behavioural interventions to be personalised and contextualised, and thus produce significant benefits. Finally, AI can help behavioural scientists to 'see the system,' by enabling the creation of more complex and dynamic models of human behaviour. While these opportunities will significantly advance behavioural science and offer great promise to improve the lives of citizens and consumers, we highlight several costs of using AI. We focus on some important environmental, social, and economic costs that are relevant to behavioural science and its application. Some of those costs involve privacy; others involve manipulation.