A new research paper from arXiv demonstrates that large language models (LLMs) significantly alter the meaning of written text, even when instructed to make only grammatical edits. A user study revealed that extensive LLM use led to a nearly 70% increase in essays that did not directly answer the topic question, with users reporting the writing felt less creative and not in their own voice. The research also found that AI-generated scientific peer reviews, which constituted 21% of reviews at a top AI conference, placed less emphasis on clarity and significance, resulting in higher scores. AI
IMPACT Highlights potential semantic drift and altered evaluation criteria due to widespread LLM use in writing and peer review.
RANK_REASON Research paper published on arXiv detailing LLM impact on text semantics and scientific reviews. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- CORE Recommender
- DagsHub
- Gotit.pub
- Hugging Face
- Influence Flower
- Large language models
- Marwa Abdulhai
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
