<p>The rise of recent AI technologies has led to significant discussion of whether it is possible to use these technologies to improve or bypass market mechanisms. However, it turns out that much of this discussion is beset by confusions. Hayek-inspired critics of AI-driven central planning are right to focus attention on the fact that the issue is not just the computation of large amounts of data, but also the fact that economic allocations depend on private or even non-existent information about individual wants and needs. However, what this response fails to note is the fact that AI may increasingly be able to predict precisely these wants and needs. This is not the end of the debate either, though, as the feasibility of central planning is not all that matters here. It also matters whether central planning is desired. Since the act of choice can express who we are as people, the answer is often no. Importantly, this desirability-focused dimension of the debate is intertwined with the epistemic dimension. Being aware of this connection allows for a more balanced assessment of when and where AI can enhance market-based allocations.</p>