This is another paper on the theme of combining information and making decisions in the face of noise and uncertainty – but the setting is quite different to those we’ve been looking at recently. Consider a food bank that receives donations of food and distributes it to those in need. The goal is to implement an automated decision making system such that when a food donation is received, the system outputs the organisation (e.g. housing authority or food pantry) that should receive it. We could hard code a set of rules, but what should they be? And who gets to decide?
A democratic solution to this would be to give each of the stakeholders a vote on every decision. In the food bank setting, identified classes of stakeholders include the donors, the recipients, the volunteers (who pick up food from the donor and deliver it to the recipient), and employees. Their votes encode their own preferences and biases, perhaps in a way that even the voters themselves couldn’t neatly codify in a set of explicit rules.