Business Standard: How AI-generated text is changing the way humans write

A Business Standard report published May 13, 2026 draws together recent research on how exposure to AI-generated content is changing the way people write — not only in professional contexts but in everyday digital communication.

The headline finding comes from a University of Southern California study: writing style diversity across scientific journals, local news articles, and social media dropped measurably after ChatGPT’s release. A parallel study from the Max Planck Institute tracked words like “delve,” “meticulous,” and “comprehend” appearing with greater frequency in spoken language after reviewing more than 740,000 hours of content — a sign that AI output patterns are being absorbed into natural speech and writing, not just into formally AI-generated text.

The article describes a specific phenomenon it calls “AI-contaminated writing”: human-authored content that has absorbed AI phrasing conventions. LinkedIn posts sound more corporate, emails are more formally structured, captions are more neatly organized. Writers are spending time removing what the report calls “awkwardly enthusiastic phrases, overly formal transitions, and formulaic wording” from AI-generated drafts, but some of that vocabulary and sentence structure is sticking in their own voice.

The paradox the article highlights is that people are simultaneously editing their language to avoid sounding like machines while their baseline vocabulary and tone are shifting closer to machine output. A Brookings survey from 2025 found that 16% of individuals already use language models for routine communication tasks — a share that has expanded the pool of AI-influenced text that human writers encounter and absorb daily.

For writers who depend on an identifiable voice — whether for editorial work, newsletters, or any kind of first-person content — the research suggests that protecting that voice now requires active maintenance rather than assumption. The default drift is toward AI-influenced patterns, and awareness of that pressure is the starting point for working against it.