ABSTRACT
Framing algorithms as taking on a role within a decision changes how we think about designing algorithms, because an important task in decisions concerns mistakes. All decisions contain the possibility of mistakes, and better decisions contain a vehicle to identify, judge, and fix mistakes. In general, managers, firms, and management researchers persistently seek to understand bad business decisions and avoid mistakes. Decisions can be unethical, unfair, bad for the long-term value creation for stakeholders, or just self-defeating. Algorithmic decisions are no different. Algorithms, whether as merely augmenting or automating human decisions, are used in important organizational decisions such as who is hired, who is fired, whether someone is deemed a terrorist, the terms offered for financing, whether an insurance company negotiates over a claim, and even how someone is sentenced. Venn diagram of types of mistakes. One circle is that the individual is blue and the other is that an individual is labeled as blue.
