5
What role do you believe artificial intelligence should play in the journal peer review process?
Results
(209 Answers)
Answer Explanations
- Supporting role onlyuser-287804perhaps AI could play a supporting role to find mistakes in language and clarity. Although it should be considered cautiously as AI might add changes that are wrong.
- Supporting role onlyuser-738101It’s difficult at this point to say, but I would like to have a human have eyes on the submission early on in the process. At the rate of change, this answer could change by next week!
- Supporting role onlyuser-193986Finding reviewers can be effectively done by AI. Reviews and decisions must be made by humans.
- Supporting role onlyuser-646537AI would be wonderful to help with general formatting and writing issues. Also REFORMATTING so submitters don't have to reformat articles for each journal. However, I don't think AI could do the bulk of checking methodology and proper science.
- Supporting role onlyuser-948023Human knowledge and actual practice cannot be overlooked when critiquing manuscripts. AI can help with language editing only. Reviewing scientific content isn't accurate by AI apps per my personal experience several times.
- Supporting role onlyuser-579540I think AI is a powerful tool, but lacks the ability to comprehend and interpret the impact of the scientific work. So, it should be used as a supportive tool
- Supporting role onlyuser-606148It's best to treat AI tools as a junior co-worker who assists the human reviewer.
- Supporting role onlyuser-426175The reasoning should be made by experts. AI can help in literature search, in filtering out style problems. Those who change (simplify) their own style, vocabulary as compared to earlier publications should be expelled from publication for a given period. (Not one journal but for all!!!)
- Supporting role onlyuser-969930Maintaining criticality in learning outcomes must be based on redl life experience
- Supporting role onlyuser-683654If AI is completely used, a lot of flaws will be left unnoticed.
- Other [Please explain ]user-524666Of course that has promise, but i think that it is premature to use it other than to be sure that all elements required of the journal are present. Someone has to do a study in which a significant number of papers are simultaneously reviewed to see if the systems (human, the "gold standard") differs from AI. Then comes the more important question. How does one know which is better. I suspect that it will take a large number of reviewers to compare with one or more AI systems.
- Supporting role onlyuser-817094There are many factors involved in peer review. AI can be used as an assistant, but it should not be relied upon to interpret data.
- Supporting role onlyuser-740158Fixing grammatical errors and flow.
- Other [Please explain ]user-320728i do not have clear idea
- Supporting role onlyuser-887652AI might be useful for copy editing. Beyond that, AI isn’t really I at all.
- Other [Please explain ]user-834781The danger of AI use by authors is not fully addressed and now this review using AI?!! After all, who will scrutinize? Sounds very scary and demotivating to remain in science.
- Major roleuser-214133AI could perform a good review, but it still requires expert supervision.
- Supporting role onlyuser-563401AI can be biased, it can help to check plagiarism or provide a report which will be evaluated by the reviewer.
- Supporting role onlyuser-781308AI can be useful in assessing the quality of language, formatting requirements, word limits, referencing style, and tables and figures in the article. AI can also be useful to examine the conformity to the research protocols. The Peer review by experts should focus on critically examining the methodology and level of evidence.
- Supporting role onlyuser-965103at it's current state. don't know about future developments.
- Supporting role onlyuser-286232Supporting in terms of coming forward with similar publications to enrich the paper and their experience.
- Supporting role onlyuser-774853AI is just as good as our understanding of the matter; it is never going to be an expert but have the expertise to bring the data together and help fill the gaps better.
- Supporting role onlyuser-41956As a 79-year-old emeritus professor, my experience with AI in scientific writing has mainly been observing students using Google search and AI-generated hallucinations. I see this creeping into manuscript submissions, typically through the conversion of a thesis into a submitted paper. Thus, I am leery of using AI- the result is only as good as the training information.
- Supporting role onlyuser-70793Artificial intelligence should only come in as a supporting cast but not to replace the main cast which, is a human being.
- No roleuser-655754I wrote "no role," but I suppose there could be some initial screening process to a) ensure no AI use in the paper, no plagiarism, b) ensure the stats check out (there is such a stats checker out there), c) check for other very low-level stuff. But basically NO I do not think AI should play a role. We're talking about generating ORIGINAL research, here. This is precisely what HUMANS can do that AI can't. AI is about rearranging and regurgitating what's already out there. This is novel and needs human oversight.
- Supporting role onlyuser-186224AI still cannot replace human intelligence on the more creative things, at least not in 5 years
- Supporting role onlyuser-104740We shouldn't give too much power to AI, it's only as good as the programmers, and they may or may not be smart enough.
- Supporting role onlyuser-767467It cannot automate a review process at this time in a complicated review system. It could definitely check math and medical facts.
- No roleuser-90744I think journal should not add AI to the peer review process. Partially for screening the plagiarism and formating it is good but factually and technicality of the peer review process has to be done by the bench scientist in my opinion.
- Supporting role onlyuser-337244The person own experience and cited research are the major part of their opinion to accept the paper. Although, AI will also boost the process of reviewing in a supportive way.
- Complete automationuser-135268I don't think it's necessary to say much. I believe this is the future.
- Supporting role onlyuser-521833To edit language for clarity. Not more than that because the models have some long way to go on subject specific areas.
- Supporting role onlyuser-881641Artificial intelligence (AI) can play a supportive but not substitutive role in the journal peer review process. Here's a breakdown of the role AI should play, along with some explanation:
1. Initial Screening & TriageAI's Role: Help editors quickly assess whether a submission meets basic journal criteria (e.g. format, plagiarism check, scope match, readability).- Why? Saves time and weeds out non-compliant or low-quality submissions early, without wasting a human reviewer’s effort.
- Tools Used: Plagiarism detectors (like Turnitin), writing quality analyzers (like Grammarly or GPT-based tools), and scope-matching systems.
2. Data Analysis VerificationAI's Role: Check the math, code, or statistics used in papers.- Why? Humans can overlook calculation errors or flawed data visualizations. AI can assist in spotting statistical inconsistencies, code errors, or data fabrications.
- Tools Used: Code auditing tools, statistical anomaly detectors, or even automated reanalysis with uploaded data/code.
📄 3. Language & Structure FeedbackAI's Role: Suggest improvements in clarity, grammar, and structure—especially useful for authors writing in a non-native language.- Why? Makes the paper more readable and polished before it reaches reviewers, improving fairness and focus on the actual science.
- Tools Used: Natural language processing tools for grammar and tone.
🕵️ 4. Reviewer MatchingAI's Role: Assist editors in finding the most qualified and unbiased reviewers based on expertise, history, and conflict-of-interest data.- Why? Finding good reviewers is time-consuming, and AI can help map author networks, specialties, and previous review history for better matches.
- Why? Saves time and weeds out non-compliant or low-quality submissions early, without wasting a human reviewer’s effort.
- Supporting role onlyuser-219698For checking plagiarism and detecting AI managed articles.
- Supporting role onlyuser-441980AI should act as a smart editorial assistant, not a replacement for human judgment. Key roles could include:
- Author and Submission Screening:
- Flag potential plagiarism, authorship issues, or conflicts of interest.
- Check for prior submissions, duplicate content, or ethical red flags.
- Flag potential plagiarism, authorship issues, or conflicts of interest.
- Quality Checks Before Review:
- Evaluate structure, grammar, formatting, and basic reporting guidelines (e.g., CONSORT, PRISMA, STROBE).
- Perform initial scans for methodological soundness, missing data, or statistical inconsistencies.
- Suggest potential reviewer matches based on topic and expertise.
- Evaluate structure, grammar, formatting, and basic reporting guidelines (e.g., CONSORT, PRISMA, STROBE).
By handling these repetitive and time-consuming tasks, AI can help reduce editorial workload, speed up triage, and allow human reviewers to focus on deeper scientific evaluation. - Author and Submission Screening:
- No roleuser-552There are researchers around me who use this. They upload the article to AI and ask me to critique it, thinking they are acting as a reviewer.
- Complete automationuser-526097If they won't pay actual academics to peer-review, then we should use AIs. No one should be expected to work for free.
- Supporting role onlyuser-332757I believe that IA should not be abused in the evaluation of papers.
- No roleuser-82487AI has many advantages, but its application to peer reviewing is still immature at this stage
- Supporting role onlyuser-573537Since AI nourishes on any material offered to it, we are incurring the risk of the informatics well-known problem of "garbage in-garbage out".
- Supporting role onlyuser-649468t o be more clear
- Supporting role onlyuser-650721They can be utilized to support but some of the nuances of a paper at this time could be missed or be influenced by biases in the learning model for the AI. They should definitely be used to eg, detect plagiarism, help identify missing citations, improve editing.
- Supporting role onlyuser-255680Screening for potential falsification of data or plagiarism.