Artificial Intelligence and Group Decision Making – Some Caveats

There is an abundant and growing literature on the ramifications of artificial intelligence (AI) for society. This paper contributes some formal observations about the influence AI might have on how people who participate in small groups interact. Within such groups (say a team, business unit, partnership or small enterprise), mutual dependence is amenable to both cooperation and conflict; some slight changes in people’s understanding of the situation can turn partners into adversaries. Using some game-theoretic examples, I argue that two frequently-heard advantages of AI – that it will enhance human productivity and wealth, and that it will support error-free and faster decisions – must be qualified in this context. Indeed, an effective algorithm, even if it delivers an accurate picture of the situation and provides every individual with proper decision aid, might uncover some sharp conflicts of interest or remove some key contextual landmarks which support human coordination. Some implications for the study and implementation of AI in organizations are briefly discussed