Cyborganisation: Machines and humans make optimal decisions together

Abstract

Business rules evolved from expert systems, a concept whereby expertise could be encoded and leveraged across an organisation. In the same way, artificial intelligence is now being used to replace human decision-makers. In both cases, the idea is to emulate and replace the decision-maker with a machine. But this approach could be considered as misguided for two reasons. Firstly, it is not possible to emulate a human with complete accuracy, and secondly, human decision-makers are fallible. Another approach is to consider how business rules and human experts can work together to maximise the expected profit of an organisation, creating a cyborganisation. There are several elements to this problem—the need to quantify the impact of different rules on the performance of the organisation, the accuracy of the human decision-maker on a case-by-case basis, and determine whether the machine and/or the human decision-maker makes the final decision. This paper considers a real example from a well-understood problem that of loan approval, but from the perspective of machine augmenting, rather than replacing, the human decision-maker. The results suggest that there are potential savings and increases in profit from this approach.