The use of artificial intelligence (AI) systems is reshaping competitive dynamics. For instance, recognizing opportunities for competitive optimization of pricing, inventory, and feature development decisions, can be done in full or in part by autonomous AI systems. While AI can generate significant business benefits, firms need to carefully consider how much decision-making latitude they are willing to transfer to AI for competitive actions. We know little, though, about when firms should prioritize AI systems that have expansive decision-making latitude versus those with limited decision-making latitude for enabling competitive actions. In this research-in-progress paper, we address this issue and present a set of propositions to serve as the basis for a new theory of AI and competitive actions.