Military organizations are progressively using AI to support decision-making, while humans continue to provide essential judgment. However, the probabilistic foundations of many AI approaches can inadvertently violate the axioms of probability when delivering answers for decision support or introduce contradictory results that induce aporia within the decision maker. Resulting mismatches between humans and AI in rational deliberation can raise challenges for decision support that relies on AI. For example, inconsistencies in rational deliberation can lead to ontic uncertainty and result in over-reliance on AI-based tools. Consequently, humans can accept an AI’s decision choice inadvertently. In this paper, we provide an overview of challenges to AI systems resulting from system attacks. We also provide a more thorough treatment of three popular Generative AI (GenAI) applications to demonstrate their inability to identify conjunction fallacies. Next, we provide an overview of quantum cognition and elucidate how it can ameliorate human decision-making supported by AI. Subsequently, we propose a concept inspired by quantum models of cognition to identify system-level interventions to improve decision-making outcomes we call a quantum cognitive alignment layer (QCAL). Finally, we propose a future research agenda to further develop a QCAL approach
Humr et al. (Wed,) studied this question.