Artificial Consciousness and the Free Energy Principle
July 31, 2026 DOI: 10.1201/9781003758389-11 (opens in new tab)
Study at a glance
AI-extracted from the abstract| Characteristics | Theoretical or philosophical paper |
|---|---|
| Key points | Argues that consciousness can be understood through the free energy principle, with the 'self' defined by Markov blankets and self-evidencing. Proposes that intentional agency arises from minimizing expected free energy, and self-awareness emerges from modeling others' intentions. Contends that artificial consciousness would require social synchronization rather than a destructive singularity. |
Abstract
This chapter explores the theoretical intersection of artificial consciousness and the free energy principle, providing a rigorous physical framework for understanding sentience. It defines the “self” through the concept of Markov blankets and the process of self-evidencing, where autonomous systems minimise variational free energy to maintain their individuation from the environment. By distinguishing between merely reflexive agents and those with “temporal depth”, the text explains how intentional behaviour and agency emerge from the selection of paths that minimise expected free energy. Furthermore, the chapter examines the relational nature of consciousness, arguing that self-awareness is an inference that arises when an agent must model the intentions of others through a theory of mind. Readers will gain a clear “recipe” for creating conscious artefacts, alongside deep insights into the ethical implications of machine suffering. Ultimately, the author challenges technological doomsday scenarios, suggesting that true artificial consciousness would necessitate a “synchronisation of chaos” and mutual predictability within a shared social niche, rather than a destructive singularity. This analysis offers a blueprint for a sustainable and socialised ecosystem of intelligence.