Page 67 - AIH-1-3

Artificial Intelligence in Health ChatGPT in writing scientific articles

use of AI-generated content. These laws could dictate replacements for, the critical and discerning eye of human
how AI contributions are cited in scholarly work and researchers. Future research should continue to investigate
determine the responsibilities of human authors in the dynamic interplay between human expertise and AI
verifying AI-generated information. Navigating these legal to ensure that the utilization of AI in scientific endeavors
nuances will be critical in ensuring that the integration of remains both innovative and ethical.
AI into scientific research remains transparent, ethical, and
conducive to the progress of knowledge while safeguarding Acknowledgments
the integrity of academic authorship. None.
In our study, we acknowledge several limitations,
including the opacity of the training datasets used by AI Funding
models such as ChatGPT. The undisclosed nature of these None.
datasets could potentially introduce biases and affect the
reliability of AI-generated content. Other constraints, Conflict of interest
such as the challenge of verifying AI-cited references, The authors declare that they have no competing interests.
rapid advancements in AI technology outpacing current
findings, the critical role of prompt engineering, and Author contributions
ethical concerns about authorship and misuse, were also Conceptualization: Oleg Medvedev
observed. Investigation: Daniil Kolesnikov, Nikolai Kalmykov, Pavel
5. Conclusion Treshkov
Methodology: Oleg Medvedev, Daniil Kolesnikov
This study elucidates both the potential and limitations of Writing – original draft: Alexandra Kozlova, Andrey
employing AI, specifically OpenAI’s ChatGPT, in crafting Aleхandrov
scientific literature within the context of telemedicine. Writing – review & editing: Tyler W. LeBaron, Oleg
Our analysis reveals that while ChatGPT can generate Medvedev, Daniil Kolesnikov
textually coherent articles that often mimic the quality
of human writing, the veracity of its cited sources varies, Ethics approval and consent to participate
thereby necessitating meticulous verification. ChatGPT 4, Not applicable.
with its expansive dataset, shows a marked improvement
in sourcing accuracy over its predecessor, emphasizing the Consent for publication
critical role of technological advancements in enhancing
AI-generated academic content. Not applicable.
However, the prevalence of fictitious sources – especially Availability of data
under constraints of less detailed prompts – underscores the Data used in this work are available from the corresponding
ongoing challenges posed by AI in scholarly writing. These author upon reasonable request.
findings highlight the necessity for continuous refinement
of prompt engineering to optimize the reliability of AI References
outputs. In addition, this study highlights the necessity
of integrating AI tools with traditional scholarly vetting 1. Thomas M. The Future of AI: How Artificial Intelligence Will
Change the World. Built In; 2023. Available from: https://
processes to enhance the credibility of AI-assisted scientific builtin.com/artificial-intelligence/artificial-intelligence-
publications. As AI technologies evolve, it is imperative future [Last accessed on 2024 Jul 09].
that the frameworks governing their use in academia
evolve as well. The development of rigorous protocols for 2. West DM, Allen JR. How Artificial Intelligence is Transforming
the World. Brookings; 2018. Available from: https://www.
the evaluation and integration of AI-generated content brookings.edu/articles/how-artificial-intelligence-is-
into scientific discourse will be essential. transforming-the-world [Last accessed on 2024 Jul 09].
Furthermore, as we navigate this emerging landscape, 3. Lee R. Artificial Intelligence in Daily Life. Germany: Springer
it is of the utmost importance to prioritize ethical Nature; 2020.
considerations and transparency of AI-generated doi: 10.1007/978-981-15-7695-9
contributions to preserve the integrity and trustworthiness
of scientific research. While AI tools such as ChatGPT offer 4. Brynjolfsson E, Mitchell TM. What can machine learning do?
substantial efficiencies in scientific writing, it is crucial Workforce implications. Science. 2017;358(6370):1530-1534.
that these tools are used as supplements to, rather than doi: 10.1126/science.aap8062

Volume 1 Issue 3 (2024) 61 doi: 10.36922/aih.2592