Viewing language models as linguistic interfaces helps in acknowledging the structural and logical nature of language. By grasping this mental model, one can avoid the issue of hallucination by not perceiving language models as all-knowing entities, but rather as tools that can either accurately correlate with reality or be entirely detached from it.