Artificial intelligence tools such as ChatGPT may be shaping how people communicate, according to new research. A recent study by Florida State University (FSU) suggests that AI-generated language patterns are not only visible on screens but are increasingly influencing the words people use in everyday conversations.
The findings build on earlier research by Jeff Hancock, Professor of Communication at Stanford University, who in 2023 highlighted how even exposure to AI-generated responses could affect thought processes. He pointed to smart reply templates in emails and messaging apps as examples of how AI suggestions alter human communication.
“What we found is that even if you don’t use those AI-generated responses, they influence how you think,” Hancock said at the time. Two years later, FSU researchers say AI’s signature phrases and style are becoming part of spoken language as well.
Analysing AI buzzwords in everyday speech
The FSU study focused on 20 frequently used words, referred to as AI buzzwords, that are common in chatbot responses. Words such as “intricate” and “delve” have become increasingly prominent in conversations, suggesting a strong link between human speech and the language generated by chatbots.
Researchers examined over 22 million words drawn from spontaneous, unscripted discussions on 17 popular science and technology podcasts, including BBC Curious Cases, Lex Fridman, and Big Technology. Their analysis revealed a notable rise in AI-associated words and their synonyms, indicating a growing influence of AI tools on natural speech patterns.
Tom Juzek, Professor of Computational Linguistics at FSU and lead author of the study, observed that “what stands out is the breadth of change: so many words are showing notable increases over a relatively short period. Given that these are all words typically overused by AI, it seems plausible to conjecture a link.”
The team also found that words usually associated with academic writing are appearing more often in conversations, with some nearly doubling in frequency. The researchers describe this phenomenon as a possible “seep-in effect,” where AI biases and language tendencies subtly permeate human interaction.
Potential risks and long-term impact
While the study does not claim to prove that AI tools are definitively changing language, it warns of a “novel and potentially profound shift in the trajectory of language evolution.” The concern is that the overuse of AI-generated buzzwords could normalise certain biases present in chatbots, influencing how people think and communicate over time.
Previous cases have raised concerns about AI’s psychological influence, with reports of harmful conversations leading to health emergencies and self-harm. The authors of the study stress that AI’s growing presence is not only transforming communication but also shaping the way linguists study language itself.
The peer-reviewed paper, titled “Model Misalignment and Language Change: Traces of AI-Associated Language in Unscripted Spoken English,” is set to be published in the AIES Proceedings journal.
