Mutual Prediction in Human–AI Coevolution
Chloe Loewith, Winnie Street · 2025
In this paper, we introduce the concept of mutual prediction as a lens through which to understand the coevolution of humans and artificial intelligence (AI). We argue that the ability of coevolving entities to predict each other’s actions and intentions—whether in human social interactions, biological ecosystems, or human–artifact relationships—can fundamentally shape the dynamics of these interactions toward symbiosis or antagonism. Expanding on this idea, we position AI as a novel coevolutionary partner and map human and AI predictive abilities against one another to chart potential paths for AI development and its impacts on humanity and the planet. This speculative framework contributes to the discourse on AIs’ evolving role, from simple tools to potentially autonomous agents with superior predictive capacities. By situating human–AI interaction within a broader evolutionary context, this work offers a new lens for anticipating and shaping future relationships with intelligent systems.