2
SciPoll 639: Using AI in Science Publication
If AI was used in the process of generating a publication should it be cited as a co-author?
Results
(241 Answers)
Answer Explanations
- Not sureuser-284533It depends on how much it was used for the whole publication. If it was used to just assist with grammar, creating tables out of already human made text, then no.
- Yesuser-964888Any sources' contribution should be acknowledged.
- Nouser-637083Not as an author - I think that should be reserved for human contributors.
But if it was used to generate material (text, images, etc) then this should definitely be stated along with an explanation that describes exactly how it was used and the extent of the AI contribution. - Nouser-397183Its not intelligent or an independent agent
- Yesuser-290459Because it is generating new ideas
- Not sureuser-28192I wouldn’t know how to include it in the authorship…
- Yesuser-352486AI is suggesting entire paragraphs and sections of the paper. By copying previous tagged work within its memory (Google) it will more than assist. These actions are affecting the actual intellectual context and content of the publication.
- Not sureuser-650602Maybe.
- Yesuser-574398To the extent that an assistant would be named as a co-author, then AI should be as well. Personally, I would tend to distrust publications that relied on AI as a co-author.
- Nouser-615884AI must always be considered a tool—an intelligent tool, but a tool like others used in research (software from simulations et al.) that must be scrutinized by the author of the work.
- Not sureuser-762237When considering the role of AI in writing and producing scientific publications, whether AI should be cited as a co-author raises important distinctions and considerations.
1. Should AI be cited as a co-author when it has contributed to the writing of scientific publications?
In this context, the answer is no. Authorship in scientific publications traditionally reflects substantial intellectual contributions that include the conception, design, execution, or interpretation of the study. AI, while undeniably a powerful tool, does not engage in intellectual creativity or original thought in the way that human authors do. Its function is to assist by processing information and performing tasks based on algorithms, without generating new insights or understanding content in a human sense.
Ethical and academic standards also require authors to take responsibility for the content and answer questions about the work. Because AI cannot assume these responsibilities or engage in scholarly discourse, it does not meet the criteria for authorship. Instead, AI should be recognized similarly to other research tools, such as statistical software or word processors, that are indispensable but do not receive authorship credit. Properly acknowledging the role of AI in the methods section or acknowledgments ensures transparency and clarifies that human authorship remains unambiguous.
2. Should AI be cited as a co-author when it was used to produce a publication?
This question is more nuanced and requires careful consideration. The extent of AI's involvement and the nature of its contributions can vary widely. If AI assists with tasks such as data analysis, proofreading, or generating suggestions, it remains a tool and should not be considered a co-author. However, as AI technology evolves and potentially takes on more significant roles in autonomous content creation or analysis, the lines may become blurred.
Authorship requires a substantial intellectual contribution, and current AI capabilities do not yet meet this criterion. However, as AI systems advance, there may be scenarios in which their contributions approach creative or original contributions, requiring a reevaluation of authorship guidelines. For now, ethical guidelines and publication policies do not recognize AI as an author. It is important to maintain transparency by clearly stating the role of AI in the research process within the publication, and to ensure that readers understand the contributions of human authors versus those facilitated by AI.
In summary, while AI should not be cited as a co-author under current standards, its role should be transparently acknowledged. As technology advances, ongoing dialogue and updates to authorship criteria may be necessary to reflect new capabilities and contributions of AI. - Not sureuser-138703not co-author but it has to be stated in a paper that analyses etc been done using AI tool
- Not sureuser-276088If all reference data corresponds and is cited in bibliography, then it does not need to be cited. However if the writing is copying word by word previous publication, a plagiarism test should be done, or the usage of AI without text change should be cited.
- Not sureuser-426136Needs to indicated perhaps as footnte
- Yesuser-475346Maybe not cited, per se, but at least acknowledged - more in terms of declaration than a "thanks for the ..." kind of acknowledgement. The declaration should include the specific program and version so this can be tracked by investigators who wish to look for patterns of reliability, etc.
- Yesuser-907425If the author doesn't come up with the argument, it should be acknowledged where the argument came from.
- Nouser-740731AI should not be used in generating a publication; it can be used only for improving the quality of language or other kinds of graphic presentation as an assistive technology.
- Not sureuser-414245If the AI discovered the results, then "yes", otherwise "no". As long as you talk about using AI as a tool to generate papers that have good English and proper significant digits etc., it is one thing. But if you ask an AI, "What is the result of all these data?" and it figures it out, then it is not the original work of the author.
- Nouser-81297It makes no sense in my opinion. At least, I believe that we may include in a paper a new small subsection with a declaration like "Part of this text has been generated with AI-name"
- Not sureuser-404499as a minimum the lead author should disclose that AI was involved in the writing process
- Nouser-905470Not necessarily as a co author but it should definitely be indicated somewhere In the paper
- Nouser-638389As I mentioned above,you are still responsible to the context. AI is just a typewriter with error free optionin.
- Nouser-52862The AI tool was used as a tool. The queries leading to the outcome are created by the human. Therefore, the human gets the credit of the outcome, much the same way as an advanced search engine is not credited when used in work. However, queries should be constructed such that any supporting resources used by the AI tool are rendered evident.
- Nouser-840552At its current stage, I believe that AI used in the process of generating a publication should not be cited as a co-author because no proper legal and institutional frameworks are set up to take account of the risk it possesses. Additionally, AI processes information based on its programming and training dataset but does not engage in original thought or creative problem-solving. Scientific research often requires nuanced interpretation and judgment, skills that AI lacks. Proper acknowledgment of the use of AI tools in the methods or acknowledgments sections can still maintain the transparency and integrity of a scientific publication rather than citing it as a co-author.
- Yesuser-543438At the very least it should be listed as a reference.
- Nouser-99098If it is not contributing new ideas and just changing style, I do not think it should be a co-author.
- Yesuser-678105Should be listed as a tool and any wording used should be attributed to AI.
- Nouser-598239No validation/verification or personal communication possible.
- Nouser-14860AI shouldn’t be use, but if used it, the author should mention in the final of paper that AI was used e how it was used
- Not sureuser-173340AI should not be the author, on the other hand, the reader should definitely know that AI was contributing to the text, perhaps even the texts written or substantially affected by AI should be clearly marked
- Not sureuser-87331I suppose it should not be cited as a co-author but it has been used in its preparation.
- Yesuser-911600Honest use of available information
- Nouser-11084AI is not a person and thus not an "author". AI is a crutch and should be cited thusly.
- Nouser-444605Do we cite all graphic softwares and google that we use as co-authors?!
- Yesuser-287804it should be noted that part of the publication was AI generated.
- Yesuser-669208It should be clearly stated that AI was used in editing/writing the text.
- Nouser-852959But it needs to be in the Acknowledgement section
- Nouser-409588It can’t be accountable for the research or the manuscript. Thus, it cannot fulfill the fourth ICMJE authorship criteria.
- Nouser-684526The human author still needs to be responsible. How do you reprimand or chastise or ban a computer if inaccurate information is published and who how will the computer be held responsible?
- Yesuser-45900It should be clearly stated that it was AI generated.
- Nouser-597118AI cannot be held accoutnable for opinions or text generation in any capacity, authorship should be restricted to individuals
- Nouser-507408As AI becomes vogue, use of any tools, and assistance including humans of the past need not become coauthors. Any identity who wpuld not claim to be authori need not be included.
If every other paper has AI as a coauthor, it will loos relevance and become a nuisance among service providers. - Not sureuser-570468It should be at least acknowledged.
- Yesuser-445218In the interest of transparency it should be noted...
- Nouser-696023If I have the suspicion that AI was used to create a manuscript or parts of it, I will reject it as a reviewer, as I have done. Science should contain logic (in the sense of natural sciences), or discuss logical chains, or seek something new; all of this cannot be present in a system trained on statistical word use.
- Nouser-596010While AI cannot be accountable for the content of the publication, and therefore doesnt qualify as an author, it must be explicitly mentioned that AI was used for transparency purposes.
- Not sureuser-445202A princípio, não! O mesmo que dizer que o word ou Excel possuem coautoria de algum texto/ IA é um mais uma ferramenta
- Yesuser-554477if not a co-author it should be stated clearly that either an image or even text was AI generated.
- Yesuser-673903Then all the mistakes can be blamed on ai.
- Nouser-544555AI cannot take responsibility so it should not be a co-author.
- Nouser-561071They should be included in the acknowledgments.
- Nouser-810586Not a sentinel being
- Yesuser-190708But under different heading of AI
- Not sureuser-67936This is an interesting philosophical question as to whether AI generates content in the same way as a human; ideally a human will synthesise and interpret knowledge, forming their own understanding. Obviously, this is not always the case and authors can simply regurgitate or paraphrase the work of others in the same way that generative AI does. Noam Chomsky has recently put forward the position that AI is a misnomer as it does not function in the same way as human intelligence does, it cannot create new ideas.
- Yesuser-987379Most of the journals have a disclaimer that you need to sign, as no AI platforms have been used to assist in writing of the manuscript.
- Nouser-123746If well established in scientific writing, i assume not to cite because authors must verify AI content and references of the information and be responsible for final output.
- Nouser-689910At this time it should be acknowledged as having been used. This might change in the future.
- Not sureuser-266855The use of the tool should be acknowledged, but it is not an author. If an entire publication can be written by Ai it would not meet the standards of new knowledge.
- Yesuser-874889As it participate in parts of the wok.
- Yesuser-97558But it need to be mentioned
- Nouser-799639When automatic tools are used in obtaining of scientific results (pipettes, centrifuges, microscopes, shakers)...they are not "credited" for the obtained results.
- Nouser-41956No, but it should be acknowledged.
- Yesuser-480376Although AI is not human, the data it uses to compose its work is derived from publications by humans.
- Not sureuser-935064Some features that involve AI, for example, correcting grammar or spelling, or offering alternative word choices, need not be cited. It seems that these uses require some level of acceptance or rejection by the author. Electronic literature searches fall in this same realm, provided that the authors use them as prompts and follow up with actually reading the original papers to verify that the papers actually make the points being claimed. To some extent, AI can be considered to be a template for organizing a flow or a coherent story. As above, so long as the authors vet every aspect of what comes from AI they could still claim it as their own. The challenge of course comes close to another age-old ethical problem known as plagerism. The question then lands in the lap of reviewers and editors -- are they able to spot such artificial constructs? This is becoming increasingly more difficult.
- Not sureuser-731405It depends on the level of participation. If he's just a proofreader, no.
- Nouser-902187Authorship credit requires assessment or evaluation of the content and responsibility for the contents validity. AI cannot take responsibility.
- Yesuser-779737People should know who wrote the paper, and exactly which parts were written by AI, so that they can make a more informed judgement about the content.
- Nouser-383159An AI tool is not an author, but a tool that needs to be disclosed, as any other
- Not sureuser-834001Still too early to decide.
- Nouser-200863We have a section called "Acknowledgement" as we do for many who help in preparing manuscripts but do not qualify for authorship.
- Nouser-330322It is not a person, it is a tool and thus being a co-author would not be appropriate.
- Nouser-802778But should be mentioned
- Nouser-940891Authors are persons.
- Nouser-642018Not as a co-author but as a reference maybe
- Nouser-809947But should give a place in bibliography
- Nouser-591482But a disclosure should be included
- Nouser-785535AI should not be used in the process
- Not sureuser-987162it is just a tool not an author
- Nouser-146796its not an author, its a tool. it should be disclosed that the tool was used however.
- Yesuser-548892The idea was not generated by you but by the machine.It's necessary for people to know that you used AI to write your document
- Yesuser-156666If anyone has contribution while writing paper then we should include his name in author name
- Not sureuser-79668De
Pends on the level
Of involvement - Nouser-111454AI often fabricates information and generates non-existent references.
- Nouser-478855Language polishing services never co-author manuscripts. They make no intellectual contribution
- Nouser-885754It should be cited as a tool that has been utilized.
- Nouser-544906The tool can be mentioned in the material method section or can be acknowledged.
- Nouser-583633All peer reviewed journals do not allow AI to be cited as an author. However most quality journals now insist on language in the article as to how AI was used in writing the article.
- Nouser-764272I think a statement in authorship contribution about how it was used is a better strategy.
- Nouser-887652Again, AI does not actually write, so it cannot be an author.
- Nouser-284769AI cannot be responsible (taken to the court) for what it creates/writes.
- Not sureuser-765807The question of whether AI should be cited as a co-author in scientific publications is an interesting one. Let’s explore the perspectives:
- Yes, as a Co-author:
- Some argue that if AI significantly contributes to the research process (e.g., data analysis, model development, or writing), it deserves recognition.
- Treating AI as a co-author acknowledges its role and promotes transparency.
- No, Not as a Co-author:
- Others believe that AI lacks consciousness, intentionality, and creativity. It operates based on algorithms and doesn’t meet the criteria for authorship.
- Citing AI as a co-author could dilute the concept of authorship.
- Considerations:
- Transparency: If AI played a substantial role, consider mentioning it in the acknowledgments or methods section.
- Community Norms: Practices may vary across fields and journals. Some explicitly address AI authorship.
In practice, there’s no consensus yet. Researchers, journals, and the scientific community need to engage in thoughtful discussions to establish guidelines - Yes, as a Co-author:
- Nouser-232098
No, it should not be cited, but its use should be acknowledged appropriately in the methods or acknowledgments section - Nouser-634057Not really, if authors use language models then to be fair they should mention this in the acknowledgment section
- Yesuser-952116It's actually not a bad idea. One would have to add the specific model version used and describe the "author" contribution in a dedicated paragraph, explaining how AI was used.