A new study suggests that the language used by artificial intelligence chatbots such as ChatGPT and Google Gemini is beginning to alter the way people speak, raising concerns over the long-term impact of machine-generated phrasing on cultural expression. The findings indicate that linguistic patterns first observed in written communication may now be spreading into everyday spoken language.
Researchers analysed hundreds of thousands of YouTube videos, podcasts and spontaneous conversations, comparing commonly used AI-generated vocabulary with alternative synonyms. According to the study, several expressions that frequently appear in chatbot responses have noticeably increased in human speech since late 2022. “We detect a measurable and abrupt increase in the use of words preferentially generated by ChatGPT, such as delve, comprehend, boast, swift, and meticulous after ChatGPT’s launch,” said Hiromu Yakura, a postdoctoral researcher at the Max Planck Institute.
See also: AI Notification Tools Lead Android 16 Upgrade Push
The results point to what the authors describe as a cultural feedback loop, in which machines initially trained on human text begin to develop their own linguistic patterns and then influence people in return. The researchers argue that this bidirectional transfer could gradually contribute to a shift in communication norms and reduce variation in spoken language over time.
The team cautioned that the trend could have broader cultural consequences, including the erosion of linguistic diversity and the loss of expression rooted in underrepresented languages. The study notes that the scale and ubiquity of LLMs could make it difficult to counteract their influence, especially as many of the dominant systems are trained predominantly in a limited number of languages and writing styles.
See also: DeepSeek Launches New Open-Source AI Models in Push to Rival Industry Leaders
While the researchers do not claim immediate harm, they suggest further study to understand how human–machine linguistic interactions may evolve and how communication norms might be preserved. The report also raises questions for policymakers and developers about maintaining cultural representation in future AI systems as their language continues to permeate daily conversation.
Source: SciAm