From Intention–Behaviour Gap to Dietary Affordances: An Active Inference Account of Sustainable Eating

Faced with the urgent challenges of climate change and environmental degradation, governments and frameworks like the EAT–Lancet Commission advocate for sustainable diets. Yet a persistent gap remains between individuals’ intentions to eat sustainably and their actual behaviour. Current approaches often attribute this gap to deficits in individual attitudes or motivation. In this paper, we challenge that view. We introduce the construct of dietary affordances to describe how opportunities for action emerge from the interplay between individuals and their environments. Crucially, we define this affordance as the synergistic product of value (goal alignment) and precision (reliability) thus implying that without sufficient reliability, high motivation (value) is mathematically insufficient to drive behaviour. Drawing on ecological psychology and contemporary active inference accounts of perception and action, we treat dietary choices not as an individual failure, but as the selection of “policies” (action sequences) that minimise risk and “expected surprise” given constraints on time, money and access. Within this framework, unsustainable eating is reframed as a rational and predictable response to a misaligned affordance field: for many, the safest, most predictable, and lowest-effort course of action is to choose foods that undermine climate goals. We argue that closing the intention–behaviour gap requires shifting the focus of interventions away from individual consumers and towards the institutions that design and govern food environments. We identify specific leverage points within the food system, including pricing, urban design and social protection, that researchers and policymakers can use to reshape the feasible behavioural policy set, ensuring that sustainable habits are not just intended but practically achievable. By reducing the expected risk and computational cost of sustainable diets, policymakers can align these options with the brain’s drive to minimize uncertainty, ensuring that sustainable habits are selected not through self-regulation, but as the natural outcome of a successfully optimized system.