How LLMs Distort Our Written Language
2026/03/18 by Marwa Abdulhai, Isadora White, Yanming Wan +4 · 16 voices · 3 citations
#cs.CL #cs.AI
paper · pdf
Abstract
Large language models (LLMs) are used by over a billion people globally, most often to assist with writing. In this work, we demonstrate that LLMs not only alter the voice and tone of human writing, but also consistently alter the intended meaning. First, we conduct a human user study to understand how people actually interact with LLMs when using them for writing. Our findings reveal that extensive LLM use led to a nearly 70% increase in essays that remained neutral in answering the topic question. Significantly more heavy LLM users reported that the writing was less creative and not in their voice. Next, using a dataset of human-written essays that was collected in 2021 before the widespread release of LLMs, we study how asking an LLM to revise the essay based on the human-written feedback in the dataset induces large changes in the resulting content and meaning. We find that even when LLMs are prompted with expert feedback and asked to only make grammar edits, they still change the text in a way that significantly alters its semantic meaning. We then examine LLM-generated text in the wild, specifically focusing on the 21% of AI-generated scientific peer reviews at a recent top AI conference. We find that LLM-generated reviews place significantly less weight on clarity and significance of the research, and assign scores that, on average, are a full point higher.These findings highlight a misalignment between the perceived benefit of AI use and an implicit, consistent effect on the semantics of human writing, motivating future work on how widespread AI writing will affect our cultural and scientific institutions.
Cited by
Discussions
- i dunno, seems bad arxiv.org/pdf/2603.18161 [bsky, 53 points, 1 comments]
- "LLMs...change text in a way that significantly alters its semantic meaning...motivating future work on how widespread AI writing will affect our cultural and scientific institutions" arxiv.org/abs/26 [bsky, 23 points, 0 comments]
- In this work, we demonstrate that LLMs not only alter the voice and tone of human writing, but also consistently alter the intended meaning. How LLMs Distort Our Written Language arxiv.org/abs/2603.18 [bsky, 7 points, 1 comments]
- Very good study!!! "How LLMs Distort Our Written Language." READ THIS STUDY. This is an A+ study. Plus there is enough information in the appendices to replicate it! arxiv.org/pdf/2603.18161 [bsky, 6 points, 0 comments]
- Pair that with this "We find that even when LLMs are prompted with expert feedback and asked to only make grammar edits, they still change the text in a way that significantly alters its semantic mean [bsky, 5 points, 1 comments]
- New from GDM: How LLMs Distort Our Written Language arxiv.org/abs/2603.18161 [bsky, 4 points, 0 comments]
- How LLMs Distort Our Written Language “LLMs [AI] not only alter the voice and tone of human writing, but also consistently alter the intended meaning” [bsky, 4 points, 0 comments]
- So research has found that LLMs distort our writing. - it edits us more than we edit ourselves. - it changes our meaning more than we would, including changing conclusions to essays - it reduces lingu [bsky, 2 points, 1 comments]
- This article, “How LLMs distort our written language,” has further convinced me that the travel book I just bought was generated by an LLM. It’s useless in that each place is described as great and te [bsky, 1 points, 0 comments]
- “our results show that using LLMs for editing leads to a large shift away from both the initial human-written drafts and the counterfactual edits that humans would have made to the same essay” (abdulh [bsky, 1 points, 0 comments]
- LLM assisted writing changes our inherent human voice. This is a feature, not a bug, but the companies who control the technology and the models. They are NOT NEUTRAL. arxiv.org/abs/2603.18161 [bsky, 0 points, 0 comments]
- arxiv.org/pdf/2603.18161 [bsky, 0 points, 0 comments]
- As for style, well, LLM output distorts writing even if the prompt is to only change style. This is because there is no semantic weight in the latent space, so they cannot separate the two. Prompt eng [bsky, 0 points, 1 comments]
- arxiv.org/abs/2603.18161 [bsky, 0 points, 0 comments]
- “In this work, we demonstrate that LLMs not only alter the voice and tone of human writing, but also consistently alter the intended meaning.” arxiv.org/abs/2603.18161 [bsky, 0 points, 0 comments]
- arxiv.org/abs/2603.18161 [bsky, 0 points, 0 comments]
Related