Competency Traps of Leveraging Third-party AI: A Simulation Approach

Organizations are increasingly leveraging third-party AI for competences previously exclusive to human capacity. However, this trend comes with costs. While third-party AI can raise organizations’ competences, it could also limit organizations’ chances to accumulate and adjust competences from directly experiencing relevant organizational activities. Through simulation experiments, we uncover two competency traps of leveraging third-party AI. First, while it may initially enhance organizations’ competences in the short term, leveraging third-party AI can hinder organizations from developing competences that could have surpassed those provided by third-party AI, leading to a performance drop. Second, even though third-party AI could provide competences that organizations themselves can never achieve, organizations may still suffer a performance decline because of the inability to update the competences provided by third-party AI dynamically. These results point to exciting lines of future research inquiry regarding how and when third-party AI should be leveraged.