A research team from the Max Planck Institute for Human Development in Berlin has conducted an in-depth analysis involving 280,000 academic YouTube videos, leading to noteworthy findings about language trends. Their study particularly highlights the usage of certain words like “meticulous,” “delve,” “realm,” and “adept.”
These terms are more commonly utilized by AI models than by human speakers. The abstract of the research paper states, “To explore whether AI has influenced human spoken communication, we transcribed and analyzed about 280,000 English-language videos of presentations, talks, and speeches from more than 20,000 YouTube channels of academic institutions.”
The research indicates a significant change in word usage patterns, particularly around terms associated with ChatGPT, following its launch. This serves as the first empirical evidence suggesting that human speakers are increasingly mimicking the language used by large language models (LLMs).
The implications of these findings raise important societal and policy-related concerns. The researchers argue that the growing influence of AI on spoken language could unintentionally diminish linguistic diversity.
Moreover, there are fears that AI language models could be misused for mass manipulation, leading to further complications in communication and understanding. As these developments unfold, the potential effects of AI on human communication warrant closer examination, not only from a linguistic perspective but also in terms of societal impacts and ethical considerations.
The study sheds light on the evolving relationship between technology and language, prompting further discourse on the balance between innovation and preservation of linguistic richness.