A recent analysis indicates that approximately one in seven biomedical research abstracts published in 2024 likely utilized artificial intelligence (AI), primarily large language models (LLMs). Out of 1.5 million abstracts indexed in PubMed, over 200,000 included vocabulary commonly associated with AI-generated content. The study, published in Science Advances, highlighted a notable rise in AI-assisted writing since the release of ChatGPT in late 2022, with researchers reporting a surge in “excess words”—style-oriented terms often unrelated to research content—in these abstracts. Notable examples of these excess words include “findings,” “crucial,” and “potential.” This lexical shift is striking, surpassing earlier trends observed during significant events like the COVID-19 pandemic. While many scholars seek to examine the impact of LLMs on scientific texting, challenges persist due to non-disclosure of AI use. Researchers express the hope that this awareness will drive more scrutiny of AI’s role in academia.
Share
Read more