The Language of AI Could Change How Humans Speak

The training data for large language models is predominantly based on written texts and scripted conversations, which may lead to a limited representation of human language. As a result, these models often fail to capture the nuances of unscripted, face-to-face conversations that are a crucial aspect of human culture¹. The widespread adoption of large language models may cause humans to encounter more AI-generated text, potentially influencing the way people communicate. This could lead to a shift in human language patterns, as individuals may start to adopt the linguistic characteristics of AI-generated content. The implications of this phenomenon extend beyond the realm of language, potentially affecting cultural expression and human interaction. This matters to practitioners because the increasing use of large language models may fundamentally alter the way humans communicate, with potential consequences for social dynamics and cultural evolution.