Code of conduct

Conflicts of Interest

What is a conflict of interest?

• A financial interest occurs when any of these points apply to the authors, recommenders, or reviewers:
  • Receive (or have received in the past four years) salaries, reimbursements, fees, fellowships, grants, or funding from an entity with financial interests that may be affected by the results of the research presented in the article.
  • Have shares or stocks in an entity with economic interests that may be affected by the results of the research presented in the article.
  • Hold patents linked to the research presented in the article
• Non-financial conflicts of interest include political, personal, religious, ideological, academic, and intellectual interests of the authors, recommenders, or reviewers that can impact the results of the research presented in the article.

The Managing Board of each thematic PCI has the right to exclude recommenders if they do not respect these rules.

PCI rules regarding conflicts of interest

Recommenders for PCI and reviewers should have no financial conflict of interest (see definition above) relating to the articles they evaluate.

Authors should have no financial conflict of interest relating to the articles they submit to PCI. Submitted preprints must, therefore, contain a section indicating that “The authors of this article declare that they have no financial conflict of interest with the content of this article.” If authors are unsure whether their article may be associated with financial conflicts of interest, they can send an email to contact [ a.t ] peercommunityin [ d.o.t ] org to ask for clarification.

Authors should have as little non-financial conflict of interest as possible relating to the articles they submit to PCI. Such conflicts of interest must be declared by the authors in the “Conflict of interest” section of their submitted article. We note, however, that some non-financial conflicts of interest may be unavoidable due to the authors’ scientific interest in the subject or the consequences that results can have on the advancement of their careers. These unavoidable non-financial conflicts of interest need not be declared, although we encourage the use of practices that limit biases arising from all non-financial conflicts of interest.

Recommenders for PCI and reviewers should have as little non-financial conflict of interest as possible relating to the articles they evaluate, although a complete absence of conflict of interest may be difficult to achieve due to the scientific interest in the subject required. For instance, recommenders for PCI and reviewers should not evaluate articles written by

• close colleagues and coworkers (in general, “close colleagues and coworkers” are considered people belonging to the same team in the last four years, people with whom they have had recurrent contacts to co-publish articles in the last four years, and have received joint funding in the last four years)
• people with whom they have conflicting relationships
• people known to have evaluated their work as a recommender in the last two years or as a reviewer in the last year
• family members, friends, or anyone for whom bias might affect the nature of their recommendation

Quality of reviews and recommendations

Reviews and recommendations should be of high quality. Reviews should be sufficiently deep and detailed for the PCI recommender handling the recommendation process to gain a full appreciation of the qualities, defects, and limitations of the article. Texts (reviews, recommendations, comments, messages to authors) will be returned to PCI recommenders and reviewers if they do not respect these rules and will remain unpublished until they do.

Communication between PCI users, members, and attendants of PCI-organized events

PCI is dedicated to providing a harassment-free experience for everyone. This includes written correspondence (e.g., reviews, recommendations, reply letters, emails) between PCI users, communication via social media, and at in-person or online events organized by or attended by PCI (e.g., meetings, seminars).

Some behaviors are specifically prohibited:

• Harassment or intimidation
• Unwelcome sexual attention, stalking (physical or virtual), or unsolicited physical contact.
• Sustained disruption or threatening speakers (verbally or physically). Speakers are asked to frame discussions as openly and inclusively as possible and to be aware of how language or images may be perceived by others
• Retaliation against an individual for reporting harassment or other unacceptable behaviors or for participating in an investigation of such a claim
• Advocating or encouraging any of the above behaviors

Any texts violating these rules may be discarded by the PCI Board.

Any participants at events organised by PCI violating these rules may be expelled by the PCI Board.

Sources:

https://improvingpsych.org/sipsinaction/code/

https://www.sortee.org/codeofconduct/

Confidential information

The names of reviewers of a recommended submission, when anonymous, cannot be shared with the public or non-involved parties. This information can only be confidentially shared upon request to journals receiving submissions of PCI-recommended preprints, as they may take PCI evaluations into account in their editorial process.

Recommenders, reviewers, managers, administrators, and data editors should not share information relative to submissions that are not recommended (e.g., not considered, rejected, or recommendation ongoing) with the public or non-involved parties.

PCI Managing Board members should not share sensitive information such as access codes (mailbox, PCI user accounts) with anyone other than PCI Board Members within their board or the PCI team. Sensitive information should be kept safe, and PCI Managing Board members should notify the PCI team if they suspect any security breach.

Use of AI

Scientists sometimes use artificial intelligence, including LLMs, to translate, correct or improve text or software code. This is similar to using traditional language correction services and poses no problem. This use is not problematic as soon as the authors review/approve/check the production.

Scientists sometimes also use AI to analyse their bibliography or their data. Data analysis using AI is not intrinsically different from other analytical methods and thus poses no problem as soon as these methods have been published and are properly cited or are original and sufficiently detailed.

Generating text, illustrations and code using AI can be problematic, because of plagiarism, AI hallucination(https://en.wikipedia.org/wiki/Hallucination_(artificial_intelligence)) (false information that is fabricated by AI) and lack of human verification.

Authorship and artificial intelligence

Artificial intelligence cannot be considered an author of an article submitted to PCI because  “All people listed as authors of the MS must meet the authorship criteria, i.e. they contributed substantially to study planning, data collection or the interpretation of results and wrote or critically revised the MS and approved its final submitted version and agree to be accountable for all aspects of the work (ICMJE 2017).”

Artificial intelligence and intellectual property

Plagiarism issues can arise when using artificial intelligence because AI tools may generate texts or illustrations identical or very similar to those found in existing sources used to train the AI models. Authors, reviewers and recommenders must ensure that no part of the manuscript has been published except for passages that are properly cited.

Using AI to generate text, code and illustrations

When AI is used to generate text and code, authors of submitted articles must

• Carefully check that there is no plagiarism, hallucinations or other errors.
• Disclose how they used AI in the section describing methods or in the acknowledgements.

Authors must not generate illustrations using AI because of plagiarism issues.

If authors do not comply with the above-mentioned rules, their submission will be rejected.

Using AI to write reviews and editorial decisions

Reviewers and recommenders must absolutely not use AI to generate their reviews and decisions. If they do so, they risk having their text rejected by the managing board of the thematic PCI and being banned from PCI. They can, however, use AI to translate, correct and improve their texts.