Organizations that aim to develop and deploy responsible AI systems must begin by considering what implicit and systemic biases reside in a technological system itself. Not doing so can result in catastrophic failures resulting from unintended consequences of the system's normal operations. The authors advise managers on how to mitigate the risk that algorithmic systems can cause organizational or social harm if left unchecked.

Human judgment relies not only on reasoning but also imagination, reflection, examination, valuation, and empathy, giving it an intrinsic moral dimension. The authors argue that just as over-reliance on navigation systems may cause us to forget how to use our innate sense of direction, over-reliance on AI-based decision support systems puts us at risk of abdicating our moral responsibilities.