The debut of ChatGPT in November 2022 stirred anticipation and apprehension within the scientific realm. The advent of this artificial intelligence (AI) tool, alongside other applications built on large language models (LLMs), prompted questions about their potential benefits and pitfalls in scientific and medical publishing. Would these AI tools aid researchers, peer reviewers, editors, and publishers in addressing the growing demands of writing, evaluating, publishing, and comprehending research? Could they alleviate challenges for non-native English speakers? Or would they inadvertently foster unethical practices like paper mills, plagiarism, and fraud?

As the landscape evolved, the quest for answers persisted. Nearly 1.5 years later, it’s prudent to reflect on the impact LLMs have had thus far. Capable of generating scientific text, from abstracts to entire papers, LLMs have proven versatile. Related tools can even produce figures and data, raising the question of whether LLMs could serve as authors.

However, most publishers, mindful of the responsibility and oversight inherent in authorship, have barred LLMs from assuming this role. While LLMs can aid in drafting and refining articles, authors bear ultimate responsibility for the content. Like humans, LLMs are susceptible to biases, omissions, and inaccuracies. For instance, when queried about the recipients of mpox, ChatGPT failed to mention men who have sex with men, highlighting blind spots in its knowledge.

Additionally, LLMs pose challenges with fictitious references, further complicating the integrity of scholarly work. Despite these concerns, the evolving landscape of LLMs continues to shape scientific publishing, prompting ongoing dialogue and adaptation within the academic community.1

You can read the full editorial here.

References

  1. Diseases, L. I. (2024). Writing and reviewing for us in AI times. Lancet. Infectious Diseases/˜the œLancet. Infectious Diseases, 24(4), 329. https://doi.org/10.1016/s1473-3099(24)00160-9

Disclaimers

  • The material in these reviews is from various public open-access sources, meant for educational and informational purposes only
  • Any personal opinions expressed are those of only the author(s) and are not intended to represent the position of any organization(s)
  • No official support by any organization(s) has been provided or should be inferred