The rise of large language models (LLMs) is changing how people communicate, both in writing and in spoken language. Through tools like chatbots, virtual assistants, and content creators, these systems are subtly shaping the way we express ourselves and interact with each other.
As more people connect with these technologies, linguistic patterns and word choices, once typical to AI, are starting to show up in human speech. This shift brings up important questions about how AI might affect language in general.
The Influence of LLMs in Communication
A study conducted by scientists from Max Planck Institute in Germany discovered clear changes in word usage patterns after systems like ChatGPT became widely available. Researchers studied 280,000 video transcripts from institutional YouTube channels and found that words like “delve”, “meticulous” and “realm” became more common.
These findings suggest a link between machine-generated language and the way humans communicate in professional or personal settings. It seems that LLMs are linguistic influencers in more ways than one.
The popularity of some words reflects how AI can shape human speech. While terms like “delve” have been smoothly adopted, others, such as “underscore”, haven’t caught on as much. This selective adoption depends on how naturally a word aligns with human conversational habits.
We could go so far as to suggest that humans increasingly imitate phrases often found in AI-edited content. Some argue, for example, that advanced vocabulary and structured phrasing, commonly seen in ChatGPT outputs, are becoming more noticeable in spoken language.
What Does the Science Say?
Howard Giles’ communication accommodation theory (CAT) claims that people naturally tweak their speech to match the language patterns of others they’re interacting with.
By frequently engaging with AI systems, we unconsciously start mimicking particular elements, such as specific word choices or sentence structures.
This isn’t a new thing. Historically, exposure to dominant communication tools has always influenced the way people speak.
What’s unique about large language models (LLMs) is their authoritative tone and precise delivery. That’s why users are more likely to adopt their language without even realizing it. These small changes then spill over into everyday spoken dialogue.
Psychology of language, a field studying how the mind processes language, shows how repeated interactions with AI tools can rewire how we think and talk. Language learning is flexible; the brain adjusts its patterns based on repeated exposure.
Over time, the brain registers these patterns and integrates them into scripted or spontaneous forms of communication. This demonstrates how external linguistic inputs introduced by AI become part of our cognitive routines.
Potential Risks of AI’s Influence on Human Communication
The expansion of large language models raises worries about shrinking linguistic diversity. As more people use AI, their speech slowly changes in a way these models communicate, which sometimes comes at the cost of unique expressions tied to specific regions or cultures.
LLMs promote language standardization, with their outputs being clear and concise, but often less varied than natural human speech.
This push toward standardization can sideline languages or dialects that weren’t heavily included in the LLMs’ training. AI communication tools sometimes struggle with handling idioms or other deeply delicate language features. These gaps make some linguistic forms less prominent over time.
The growing preference for LLM-generated language encourages uniformity and narrows the range of words we use.
Because of this, the dynamic systems of language shaped by cultural influences may lose their depth. Over time, informal writing and natural stylistic quirks could fade as AI-created patterns become more embedded in daily and professional communication. This reliance on standardized expressions reduces the richness that makes human communication exceptional.
In Summary
The analysis shows how LLMs are shaping human spoken communication, with noticeable changes in speech patterns influenced by AI outputs. For example, the rising use of words like “delve” and “meticulous” reflects a connection between human language and AI-generated phrasing.
This suggests that AI-driven clarity and structure are fusing into how people naturally communicate.
Still, there are concerns about losing linguistic diversity and the risk of making language too uniform.
It’s important for future research to look into how these changes will affect language. We must try to find a way to protect cultural and linguistic originality while embracing new technologies.
