Advantages

Making a linguistically coherent text from bullet points or short notes collected from different references while performing a literature review

Sorting and managing references

Performing study design, analysis of the study results and drafting them into text, which is thought to help writers overcome “Blank-sheet syndrome.”

ChatGPT, with the supervision of a human author, significantly shortened the manuscript writing time, reaching up to one hour for a complete manuscript draft

It could help with different manuscript preparation steps, including formatting, linguistic correction and translation for non-English-speaking authors

It could rapidly analyze and summarize data from many articles, which is helpful while performing a literature search and review

Overall, it helps improve the writing quality by detecting areas of inconsistencies and potential errors

Plagiarism detection could be performed using these tools

Disadvantages and worries

Text generated by ChatGPT could be a mix of accurate and fabricated data

Sometimes, the references suggested by ChatGPT are either wrong or non-existent

The text generated to answer a particular question is not necessarily accurate, and the answer could change if the question were asked multiple times

Controversy regarding the intellectual property rights if the text was generated using ChatGPT, do they belong to the human authors or the algorithm?

ChatGPT does not necessarily follow the journal instructions if asked to formulate a manuscript according to specific journal authors’ instructions

It is worth noting that OpenAI incorporated a disclaimer in all chat windows that AI or ChatGPT could offer incorrect and biased information

Possible unethical use of these tools (however, still not proven), such as its use to produce vast amounts of abstracts by organizations such as paper mills or even the production of fabricated research articles

AI assisting tools could help researchers to increase their research production; however, this comes without an actual increase in their real experience. This could make publication numbers less trustworthy as a parameter for promotion or hiring academics

If fabrication or plagiarism were detected within a text or manuscript generated with the assistance of AI tools, who would bear the responsibility and the consequences?

Future perspectives

Changing the regulations and policies needed for evaluating scientific manuscripts submitted to journals and abstracts submitted for conferences, including using AI-generated text detectors

AI could autonomously publish different manuscript types (commentaries, editorials and reviews)

Implementation of new ethical roles, including clear disclosure regarding using AI-assistive tools during preparing manuscripts and at which stage it was used. Furthermore, journals should also disclose their use of AI-generated text detection tools

Other ethical issues that are not yet resolved, such as whether ChatGPT or any other AI-based tools should be listed as an author or not; furthermore, these tools are currently free of charge, and these could soon be paid tools, which could further increase the gap between researchers from high- and low-income countries as well as between senior and junior researchers

AI-detector tools investment will flourish, the same as what happened with plagiarism checkers

Theoretical and practical implications (based on the current review results)

For researchers

• -

  Judicious utilization of AI-related tools in various aspects of scientific writing (such as translation and editing) could be beneficial

• -

  Preserving the research community integrity is the responsibility of each researcher

• -

  Honest reporting of using these tools during manuscript preparation is mandatory

For decision-makers in higher education institutions and medical research

• -

  Institution-specific regulations in the light of the reported advantages, disadvantages and ethical issues should be developed and implemented to regulate the use of these tools during various stages of scientific research

• -

  Educational curriculums could be introduced to inform researchers of the benefits and harms of using these tools. Furthermore, it advises on the best practices and the trustable tools to be used

• -

  Using these tools during critical thinking for proper decision-making regarding healthcare-related decisions

• -

  Instead of ignoring or prohibiting these tools altogether, there should be beneficial controlled introduction and involvement during various processes of education, training, managing patients (adjusting drug doses, remote monitoring, symptoms development checks) and improving medical research (help in designing studies, recruiting patients and documentation)