This research investigates how Large Language Models are reducing linguistic diversity across written communications, with significant implications for identity expression and linguistic analysis.

Key Findings:

• LLMs create more homogeneous language patterns regardless of the writer's background
• AI-assisted writing reduces distinctive linguistic markers that reveal personal and cultural identity
• The homogenization effect compromises applications in fields like healthcare, marketing, and user experience design
• As LLM usage increases, our ability to extract valuable insights from linguistic analysis decreases

Why This Matters: For linguistic researchers and professionals, this homogenization threatens the rich data traditionally available from language analysis. Organizations relying on linguistic cues for audience understanding may need to develop new methodologies as LLM usage becomes more widespread.

The Shrinking Landscape of Linguistic Diversity in the Age of Large Language Models