The Silent Meddling of Algorithms
Carina Prunkl · 2024
Abstract Suppose AI is used in a company’s call centre to pair incoming customers with the most suitable employee—the goal is to maximise customer satisfaction. This seems to be a good result for the company and a good result for the customer. But the author suggests this process is not as unproblematic as it appears. We need to ask why certain inferences are being drawn by the AI algorithm—and whether these inferences are based on stereotypes and historical injustice.