Ethical policies for reviewers

Considering the evolving capabilities of generative artificial intelligence (GenAI), including large language models (LLMs) and AI chatbots such as ChatGPT, this policy outlines acceptable and unacceptable uses of these tools by reviewers.

Why this policy matters

Generative AI tools can offer real benefits when used responsibly, such as improving clarity, readability and accessibility of peer review reports. They can help reviewers communicate feedback more effectively and overcome language barriers.

However, these tools are not infallible. They can produce misleading or fabricated content, lack accountability, and cannot guarantee confidentiality. Uploading manuscript material to GenAI platforms may expose sensitive data to third parties, breaching author rights and privacy laws. Misuse of AI in peer review risks damaging trust in the process and undermining research integrity.

This policy aims to balance these considerations, safeguarding confidentiality, accuracy, and fairness while allowing transparent use of AI for language improvement.

Acceptable use

Reviewers may use GenAI tools only to assist with improving the written quality of their peer review reports. This includes:

• Checking grammar, spelling, punctuation, and formatting
• Enhancing clarity, flow, and structure of the language
• Translating a completed review from the reviewer’s first language into English

Reviewers should disclose any use of GenAI tools for language editing and translation. Reviewers (including co-reviewers) remain fully responsible for the contents of their reviews and for ensuring their accuracy, integrity, and originality.

Unacceptable use

Reviewers must not:

• Upload any part of a manuscript under review, including text, figures, tables, supplementary files, or associated communications, to GenAI tools
• Use Gen AI tools to analyse, summarise, interpret, or evaluate the manuscript or its scientific content
• Use GenAI tools to generate references or citations
• Use GenAI tools to write any portion of the review report beyond language refinement

If a reviewer misuses GenAI tools, we may disregard some or all of their review.

Reviewer responsibility

By accepting a review invitation, reviewers agree to:

• Adhere to IOP Publishing’s ethical standards, including confidentiality, conflict of interest disclosure, and maintaining anonymity
• Ensure that the manuscript under review is not shared with any third party, including GenAI tools
• Take full responsibility for the content of their review, whether conducted individually or as part of a co-review

GenAI tools are not expert peers. They lack higher-level reasoning and critical thinking, are prone to generating false or misleading content (“hallucinations”), and cannot take responsibility for the material they produce. They also lack the legal personality to sign publishing agreements or licenses and cannot be held accountable for ethical compliance.

To uphold impartiality, you should consider any potential conflict of interest before agreeing to review and should contact the editorial office in the following instances:

• You are in direct competition with the authors
• You are a co-worker or collaborator with one of the authors
• You are in a position to exploit the authors’ work (commercially or otherwise)
• You may be legally prohibited due to national sanctions
• You are in a position which prevents you from giving an objective opinion of the work.

Minor conflicts do not disqualify you from reporting on an article, but will be taken into account when considering the reviewers’ recommendations. Major conflicts of interest (especially relating to a financial commercial interest >£5000/year) do disqualify you. You should act within the spirit of the Nolan principles of public life.

If you are unable to act as a reviewer due to a conflict of interest, we will select an alternative reviewer.

If the journal is double-anonymous you may not be sure if you have a conflict of interest. If you suspect there may be a reason you should not act as a reviewer, please contact the editorial office who will be able to investigate and advise.

Reviewer names are kept strictly confidential. Reviewer identities may only be disclosed to journal Editorial Board members, who are also instructed to maintain confidentiality. You should not disclose your identity to the authors, including sending reports directly to the authors.

Information and ideas obtained whilst acting as a reviewer must be kept confidential and not used for competitive advantage.

We also ask that you do not discuss the papers you have reviewed with colleagues unless they have been published.

Reviewers should judge objectively the quality of the research reported, give fair, frank and constructive criticism and refrain from personal criticism of the authors. Reviewers’ judgements should be explained and supported so that authors can understand the basis of the comments and judgements.

Reviewers should inform the journal if they are unable to review a paper or can only do so with some delay. Reviewers should not delay the peer review process unnecessarily, either deliberately or inadvertently.

Reviewers are expected to point out relevant work that has not been cited, and use citations to explain where elements of the work have been previously reported. When writing a report, reviewers should justify any literature references suggested for inclusion in the work.

Citations should add value, and should not be unfairly biased towards an individual, group or organisation. Please note that the Editor reserves the right to challenge excessive citation suggestions, especially to the reviewer’s own work. The practice of including superfluous references, including to the reviewer’s own work, to promote and inflate citation scores is unethical. The Editor reserves the right to exclude citation suggestions from reports if these are considered to be potential acts of citation manipulation, and/or to protect reviewers’ anonymity.

Reviewers should report any suspicions of misconduct to the journal staff for investigation. This includes, but is not limited to, suspicions of:

• Plagiarism
• Duplicate publication
• Parallel submission
• Data fabrication / falsification
• Image manipulation
• Incorrect authorship
• Author conflict of interest
• Unethical research practices
• Content that could be considered offensive

We follow the COPE guidelines on responding to whistleblowers, which includes protecting your anonymity.