Artificial intelligence is reshaping not only how machines communicate but how people write and speak, according to two new academic studies. Researchers at the University of Southern California and the Max-Planck Institute for Human Development found that tools like ChatGPT are driving a measurable decline in stylistic variety across scientific journals, local news, social media, and everyday conversation.
The USC study, led by professor Morteza Dehghani, analyzed writing samples from multiple domains and detected a sharp drop in diversity of sentence structure and vocabulary after ChatGPT's launch. Meanwhile, the Max-Planck team reviewed 740,249 hours of content and found that words favored by ChatGPT—such as "delve," "meticulous," "boast," and "comprehend"—are appearing more frequently in daily speech.
Dehghani explained the phenomenon as a feedback loop: "People get used to this idealized, very predictable form of language, and even people who are not using it, in order to have that sense of powerful, influential writing, they start writing more like LLMs." This suggests the influence of AI extends beyond direct users to the broader culture.
Critics argue the homogenization comes at a cost. Alex Mahadevan, chief AI instructor at the Poynter Institute for Media Studies, described AI-generated writing as "soulless" and "mediocre." The findings raise questions about whether the pursuit of efficiency and polish in communication is eroding the creativity and unpredictability that define human expression.
Some researchers caution that the studies may conflate correlation with causation, as other factors—such as shifting media norms or the rise of short-form content—could also contribute to the observed linguistic shifts. The long-term cultural impact remains uncertain.