A new survey finds that 32% of small businesses now use artificial intelligence for outreach and customer service, while 16% of individuals rely on large language models for social media or general communication. The findings amplify concerns that AI has moved from a tool of convenience to a force reshaping everyday expression. As models become a default aid for drafting emails, posts, and promotional copy, researchers and linguists say a “perfect,” predictable style is spreading across public and private communication. Even people who do not directly use such tools are imitating the model-driven register to make their text sound more professional or influential. This suggests the influence of AI-generated phrasing extends beyond the users of these systems.
Academic and media language appear to be shifting in tandem. Studies from the University of Southern California report a sharp decline in stylistic diversity in scientific papers, news writing, and social media posts following the emergence of AI-based tools like ChatGPT. At the same time, researchers at the Max Planck Institute for Human Development have found that words commonly favored by these systems—such as “delve,” “meticulous,” “boast,” and “comprehend”—are appearing more frequently in everyday conversation. Linguists note it is often difficult to distinguish AI-generated text from human writing, a sign of growing uniformity in tone and structure.
The dynamics behind the shift are described as a standardization effect. AI tools nudge users toward recurring sentence structures, transitions, and a familiar set of words. Those patterns increasingly show up across contexts as people unconsciously gravitate toward what reads as polished and effective. The trend is also described as a pursuit of a particular register that smooths away idiosyncrasies and voice in favor of consistent output that feels optimized for broad audiences. “The trend is a pursuit of ‘ChatGPT level polish,’ which dulls authentic individual voices and encourages a style referred to as the ‘LinkedIn average,’” said Emily Bender, a linguist at the University of Washington.
Model-favored vocabulary
Researchers tracking word frequency say the rise of model-favored vocabulary—exemplified by “delve” and “comprehend”—illustrates how quickly stylistic habits can propagate once they are reinforced by recommendation loops and widely used drafting aids. In practice, the resulting prose is often described as grammatically correct but lacking the sense of art or unique engagement that has traditionally marked human expression. The effect is not limited to those who rely on AI for first drafts. As readers become accustomed to this style, many writers adopt similar phrasing to emulate what they perceive as high-impact communication. This push toward uniformity extends across professional settings, where standardized language promises fewer errors and a consistent tone, and into personal exchanges, where users mirror what they encounter in feeds, forums, and inboxes.
The pattern, sometimes summarized as a shift toward a “LinkedIn average,” has prompted warnings from linguists and communication researchers that an incremental unification of expression could crowd out regionalisms, specialized jargon, and individual voice.