Auditory verbal hallucinations (AVH) - experienced as voice hearing independent of a corresponding external sound source - are a cardinal symptom of psychosis. Approximately 6-13% of healthy individuals also experience voice hearing. Despite numerous attempts to explain the neurofunctional mechanisms underlying AVH, they remain notoriously unexplained. However, evidence relates AVH to mechanistic changes in the forward model. This review synthesizes behavioral and neuroimaging studies exploring the central role of cerebellar circuitry in the forward model, with a particular focus on non-verbal and verbal auditory feedback. It confirms that erratic prediction of sensory consequences in voice and sound production is linked to impaired cerebellar function, which initiates AVH and affects higher-level cognitive functions. We propose new research directions linking the forward model to voice and sound feedback processing. We consider this review as a starting point for mapping mechanisms of the forward model to neurocognitive mechanisms underlying AVH.