One of the biggest science stories in 2023 according to the newspaper Guardian was the mainstream use of Artificial Intelligence (AI) in particular the use of large language models such as ChatGpT . AI will undoubtedly become a transformative force in the field of musculoskeletal science, influencing various aspects, including scientific article writing. The integration of AI brings about both benefits and risks that researchers and authors need to carefully consider. What are the potential benefits of using AI when drafting a scientific article? AI is generally fast and efficient unlike humans. AI tools can significantly enhance the efficiency of literature reviews and data analysis and can quickly sift through vast amounts of information and extract key findings. In addition to speed and efficiency AI-powered language models, such as natural language processing, may assist in improving the overall writing quality of scientific articles. They can suggest grammatical corrections, enhance readability, and offer constructive feedback on writing style. Additionally, AI algorithms are excellent in analyzing complex datasets and identifying patterns. This capability potentially aids researchers in uncovering hidden correlations within medical data, leading to more robust scientific insights. Lastly, automated citation and referencing can streamline the process of citing relevant studies and accurate referencing which may reduce the likelihood of errors.
However, there are multiple risks when using AI and the authors must critically evaluate and address potential errors such as bias. If the input data used to train AI models is biased, it can perpetuate and even exacerbate existing biases in medical research. Also depending too heavily on AI tools may lead to a decreased emphasis on critical thinking and human expertise. Researchers must maintain a balance, using AI as a supportive tool rather than a substitute for human judgment and interpretation. Lastly handling sensitive medical data raises concerns about security and privacy. Authors need to be vigilant about protecting patient information and ensuring compliance with ethical guidelines when utilizing AI in research. Further aspects of pros and cons of artificial intelligence are summarized in an article by Salvagno et al “Can artificial intelligence help for scientific writing?”
As an example I have asked the freely available version of ChatGPT 3.5 what the preferred imaging modality for the diagnosis of Chronic Non-Bacterial Osteomyelitis is and asked for appropriate referencing. The listed modalities (MRI scans, Xrays and Nuclear Medicine Bone scans) are however incomplete and omit the role of CT or PET-CT scans particularly when examining anterior chest wall or the whole body. Additionally, the output of this ChatGPT version included several older references from over 20 years ago which includes outdated information such as x-ray examination to assess inflammatory activity, and omits the most recent papers which have been published after September 2021. In summary for this musculoskeletal question ChatGPT was not very helpful or accurate.
In conclusion, the integration of AI in writing scientific articles in medicine may offer substantial benefits in terms of efficiency, data analysis, and language enhancement. However, authors must navigate the risks associated with bias, overreliance on AI, and security and privacy concerns to ensure the responsible and ethical use of these powerful tools. With increasing use, we will likely need a consensus on how to regulate the use of chatbots in scientific writing and journals or societies will have to issue specific guidance on the use of AI.
The author of this article acknowledges the use of ChatGpT 3.5 in writing this article which reduced the writing time by approximately 30%.