The language of science is constantly evolving, with new words and phrases entering common usage as scientific research progresses. Recently, the introduction of ChatGPT – an artificial intelligence tool – has revolutionized the way scientists write their papers. Since its launch in November 2022, scientists have been utilizing Large Language Models to edit, revise, and even generate scientific content.
Researchers led by Dmitry Kobak at the Hertie Institute for AI in Brain Health have analyzed the impact of AI systems on scientific literature. By examining over 14 million scientific paper abstracts from the PubMed biomedical database since 2010, they discovered that Large Language Models are reshaping scientific discourse on an unprecedented scale. The team identified a significant increase in words related to writing style, such as “delves,” “crucial,” and “important,” which are believed to be favored by ChatGPT.
This surge in style-related vocabulary in 2024 serves as a marker for the widespread usage of ChatGPT in scientific paper writing. The researchers estimate that at least 10% of papers on PubMed in 2024 were influenced by Large Language Models, equating to around 150 thousand papers per year. Interestingly, AI assistance was more prevalent in papers from non-English speaking countries, indicating a potential leveling of the playing field in scientific writing.
The impact of AI on scientific writing is profound, surpassing even the vocabulary changes brought about by the Covid-19 pandemic. As scientists and publishers navigate this new landscape, it is essential to understand and address the implications of AI assistance in scientific research. By acknowledging these trends, the scientific community can establish guidelines for the ethical use of Large Language Models to benefit scientists, publishers, and society as a whole.
The rapid adoption of AI in scientific writing underscores the need for swift adaptation to keep pace with evolving technologies. Similar challenges are likely to be faced in other areas of publishing, highlighting the importance of proactive measures to harness the potential of AI while upholding the integrity of scientific communication. As we embrace the era of AI-assisted scientific writing, a collaborative effort is required to shape the future of scientific discourse and ensure that advancements in technology serve the greater good of society.
