The article discusses the rhetorical device known as 'negative parallelism,' exemplified by phrases like 'It’s not X; it’s Y.' This construction appears frequently in AI-generated text, including corporate communications and literature. A report by Barron’s indicates that the use of this device in corporate communications has increased significantly from 2023 to 2025. Researchers at Pangram estimate that such sentences are three times more common in AI writing compared to human writing. The article also explores the reasons behind the prevalence of this construction in AI models, suggesting that it may stem from the patterns in the training data and the reinforcement learning process used to refine these models. Experts note that while this rhetorical device can enhance clarity, its overuse may lead to formulaic writing. The persistence of negative parallelism raises concerns about the potential for AI-generated text to increasingly resemble itself, complicating efforts to improve AI writing quality.
Why this rating? · 1 signal
Signals flagged in the original
- headline asserts a conclusion / scare-quotes
Provisional estimate — refines shortly Full breakdown ↓
Analysis of AI Writing Patterns in Language Models
The article examines the rhetorical device known as 'negative parallelism' in AI-generated writing, highlighting its increased prevalence in corporate communications and literature. Researchers attribute this trend to patterns in training data and the reinforcement learning process used in AI models. While this construction can enhance clarity, its overuse may lead to formulaic writing.
Compare the coverage
No note attached
on this article.
Bias Analysis
Bias Indicators Removed
- ✕ headline asserts a conclusion / scare-quotes
Original vs. Neutral
It’s Not Just Annoying, It’s Inescapable
Analysis of AI Writing Patterns in Language Models