Page 10 - IJSA, Vol. 6, No 1, 2023

International Journal of Science Annals, Vol. 6, No. 1, 2023
рrint ISSN: 2617-2682; online ISSN: 2707-3637; DOI:10.26697/ijsa

However, there is a precedent of ChatGPT having a ChatBot. In doing so, the question posed to the ChatBot
profile in Scopus (ChatGPT, n. d.), as well as papers and the answer received from the ChatBot should be
published by prestigious international publishers in clearly stated.
which ChatGPT is listed as an author (O’Connor & In our opinion, information about the use of ChatBot
ChatGPT, 2022). should necessarily be reflected in the methodology with
Also noteworthy is the book “Impromptu: Amplifying a correct indication of which ChatBot was used by the
Our Humanity through AI”, in which GPT-4 writes: “I author, where and to what extent. The name of ChatBot
would like to thank Reid Hoffman for inviting me to co- and its characteristics should be specified in the
author this book with him”. Please note that Reid References list.
Hoffman, a leader in the field of AI, states on the title Our recommendation is also based on the fact that in the
page “By Reid Hoffman with GPT-4” (Hoffman with near future it will probably be impossible to hide the
GPT-4, 2023). involvement of ChatBots in the writing of a scientific
There is one case in the literature where ChatGPT has paper. ChatBots-creating companies will start using
answered negatively to the question of whether it meets something like a “watermarking” on the bot’s output to
all of the International Committee of Medical Journal make plagiarism easier to spot. The San Francisco-based
Editors (ICMJE) criteria for authorship – “ChatGPT can company OpenAI, which created ChatGPT, has already
assist in the drafting or revising of a work, but it cannot announced this. OpenAI guest researcher Scott Aaronson
fulfill all of the ICMJE criteria for authorship” said that “the technology would work by subtly tweaking
(Anderson, 2023). the specific choice of words selected by ChatGPT, …, in
Perhaps it is a question of specific criteria for authorship, a way that wouldn’t be noticeable to a reader, but would
rather than ChatGPT’s refusal to acknowledge its role in be statistically predictable to anyone looking for signs of
writing. In any case, we have not received a clear answer machine-generated text” (Hern, 2022).
to this question. Therefore, the answer should be sought So there is a good chance that if you try to pretend to be
in the aspect of ethics, as well as the willingness of the the author of text written by a ChatBot, you may be
person to recognize the authorship of ChatGPT or not. detected. Turnitin has already begun work on developing
Todd Carpenter conducted a ChatGPT survey on the an AI-based text detection tool (Chechitelli, 2023).
impact of AI on science communication. Specifically, he In early April 2023, the American Psychological
asked about the ethics for an author of using AI in Association (APA) website published information with
developing a scholarly paper. As ChatGPT learned from guidelines for quoting and reproducing text generated by
the response, ethics “depends on the specific context and ChatBots (McAdoo, 2023).
the expectations of the research community in which the We recommend that Authors of our Journal use these
article will be published” (Carpenter, 2023). standards when preparing a manuscript and citing text
ChatGPT itself sees no ethical problems with the use of generated by ChatBots.
AI in scientific writing. However, it notes that authors It is important to note the statement of the Committee of
must “clearly state this in the article and provide Publication Ethics (COPE). On its website, the
appropriate credit to the AI program” (Carpenter, 2023). Committee has published its official position on
Springer Nature and Taylor & Francis Publishers suggest authorship and the use of AI tools (COPE Council, 2021;
that AI contributions should be reflected in the methods COPE, 30 January 2023; COPE, 13 February 2023;
or acknowledgements section, rather than being listed as COPE, 23 February 2023; Watson & Stiglic, 2023). Also
an author (Stokel-Walker, 2023). a number of papers on using AI for scientific writing
This position is justified by the important characteristic (Çalli & Çalli, 2022; Dans, 2019; Dimitriadou & Lanitis,
of authorship – responsibility for publication. 2023; Farahani, 2023; Singh & Sood, 2022).
In this context, it should be noted that it is known that AI Today, COPE is virtually the only organization in the
has convincingly described the results of studies scientific world that promotes ethical principles in
(specifying the organizations that conducted them and scientific publishing. COPE Council members warn that
the quantitative indicators). However, when clarifying the increasing role of AI in research writing “has
the information, he could not confirm it with any sources significant implications for research integrity and the
and apologized for the error and confusion in his need for improved means and tools to detect fraudulent
statement (Davis, 2023). research” (COPE, 23 March 2023).
These facts point to the need for caution and responsible This is a matter of concern for those scientific publishers
use of information obtained from AI. It is important to who conduct their activities responsibly and put into
remember that human remains responsible and practice the principles of scientific publishing ethics and
accountable for copyright infringement. the COPE standards.
If someone claims undivided authorship, he/she should The IJSA is a full member of the COPE (COPE, n.d.).
objectively, based on facts, state the role of ChatBot in Thanks to this, the members of the IJSA Editorial Board
the scientific publication, claim full responsibility for the were able to participate online in events dedicated to the
content of his/her manuscript and the result, including the discussion of this topical issue (COPE, 23 March 2023).
parts created by ChatBots, as well as the degree of
originality of his/her publication. Perhaps there is no Conclusions
shame in stating that the research design, data collection, We started our Editorial with a warning: this paper was
or statistical analysis was done using a particular not written by a ChatBot and is intended for humans.

8