Strategic interactions under AI-driven demand forecast: Information sharing and supply chain coordination

An accurate demand forecast by artificial intelligence (AI) reshapes information asymmetry and strategic interactions in a decentralized supply chain, where the observability of real-time market fluctuation originates from the AI-driven demand forecast. This paper constructs an analytical model of the supply chain to examine the retailer's demand prediction strategy and the manufacturer's information sharing strategy. Research shows that the retailer prefers AI-driven demand prediction when the unit computing power cost is low, while achieving information sharing can lead to a win-win situation. Although the accurate demand forecast improves the manufacturer's payoff, higher cost allocation of AI adoption may also deter information sharing as the unit computing power cost increases. In contrast, if the retailer adopts traditional demand prediction, they can only reach an information-sharing agreement through a subsidy contract. Once voluntary sharing is reached, such an agreement can constantly be reached under AI-driven demand prediction but fails under traditional demand prediction due to the absence of the subsidy cannibalization effect. Furthermore, it is through the subsidy contract that the demand distortion can be prevented from harming the manufacturer's payoff when the retailer inflates forecasting demand. This study provides enterprise managers with a better understanding of tradeoffs between the costs and benefits of AI adoption and makes prescriptions regarding how information sharing can be achieved for supply chain coordination.