ChatGPT's emergence is reshaping conversational dynamics in Germany
A research paper published by a group of scientists from the Max Planck Institute for Human Development in Germany has shed light on the impact of large language models (LLMs) like ChatGPT on human speech, suggesting that these models could be contributing to a noticeable shift in vocabulary and communication style.
The researchers imply that taking vernacular cues from AI systems with no understanding of their own language usage could be problematic, as there could be long-term ramifications of AI-human interactions on how both use language. This could potentially create a closed cultural feedback loop between humans and machines.
Since the debut of ChatGPT around late 2022, humans have increasingly adopted what researchers call "ChatGPT-style" vocabulary. Words such as "delve," "meticulous," "adept," "realm," and "navigate"—once relatively rare in everyday speech—have surged in frequency by up to 51% across spoken and written communication. This demonstrates a cultural feedback loop where AI, originally trained on human language, is now feeding back into and reshaping how people communicate.
The influence extends beyond just word choice. The tone of communication has shifted to a more polished, formal, and emotionally neutral style that mirrors AI-generated text, potentially leading to more uniform and less personally distinctive speech. While AI-inspired language can sound confident and professional, it may simultaneously suppress individual quirks, regional dialects, and the natural diversity that characterizes human languages.
This homogenization process risks eroding linguistic diversity by promoting a standardized mode of expression at the expense of unique cultural and personal speech patterns. The implications are significant for cultural identity since language is deeply tied to culture. As more people incorporate AI-influenced phrases and styles, distinctive dialects and expressive forms might diminish over time.
However, experts emphasize that this transformation is ongoing, and society still controls how to balance the efficient, “perfectly sophisticated” AI-influenced voice with the rich, chaotic beauty of human linguistic diversity.
The researchers also express concern that an AI/human linguistic discourse could lead to cultural homogeneity with sufficient time and reach. They suggest that if AI systems disproportionately favor specific cultural traits, they may accelerate the erosion of cultural diversity.
Despite the concerns, the paper does not pass judgment on whether the shift in language usage is good or bad but raises the question of whether large language models are shaping human culture. The research is a non-peer-reviewed preprint copy, and it does not discuss the potential impact of AI on the evolution of human speech patterns or the diversity of human cultures.
As AI continues to evolve and permeate various aspects of our lives, it is essential to consider the potential consequences on our linguistic and cultural diversity. The research serves as a reminder for society to actively balance the benefits of AI-driven language usage with the preservation of our unique linguistic and cultural identities.
The researchers' findings indicate that the rise of AI-influenced vocabulary and communication styles, such as those seen in ChatGPT, could potentially lead to a homogenization of language, particularly through the suppression of individual quirks, regional dialects, and linguistic diversity. Furthermore, the increased use of AI-inspired phrases and styles in human communication could create a feedback loop with AI systems, potentially escalating the erosion of cultural diversity over time.
