Loss aversion is a theory of behavioral economics that predicts that people won’t try as hard when their competitors are doing better. Recent research has now validated this effect in systems where people compete against machines. The researchers found that people found themselves less competent against a fast, competitive robot, even though there’s no direct interaction. When people rated the machine as performing better, they also rated its competence higher, its likability lower and their own competence lower. For employers, this should be an important factor in the design of any AI system where humans can potentially feel the machine is competing against them.
The implication for human-centered AI design is that, in situations where humans and machines are working alongside each other, the competitive impact on the human needs to be taken into account. According to Guy Hoffman, assistant professor in the Sibley School of Mechanical and Aerospace Engineering, “while it may be tempting to design such robots for optimal productivity, engineers and managers need to take into consideration how the robots’ performance may affect the human workers’ effort and attitudes toward the robot and even toward themselves.”
Machines can affect human competitive instincts in different ways, most notably when a machine enhances human capability. In a call center an AI can listen to calls and track voice tonation. The AI nudges human operators to slow down, be patient, sound less robotic and check for understanding. Not only does the deployment increase customer satisfaction and loyalty, it also increases employee performance and retention. In one deployment we studied, this wasn’t because agents felt better about the interaction with a client in need (although that does happen), it was because the AI changed how agents competed with their peers. The AI found patterns between an agent’s emotional skill and call satisfaction measures and was able to prove to an agent’s manager that greater human emotional intelligence reduces cost and risk. The highest performing agents liked the AI because, for the first time, there was human-centered data about what made them better agents.
Loss aversion is a feature of human psychology and a potent driver of decision making. Keeping humans feeling like they are “still in the game” may mean a slight drop in the efficiency of the machine but the human-machine system will be better overall. Similarly, using machines to enhance human capability can increase performance by providing measures for previously immeasurable factors.
