JOSS Policies

Animal research policy

In the exceptional case a JOSS submission contains original data on animal research, the corresponding author must confirm that the data was collected in accordance with the latest guidelines and applicable regulations. The manuscript must include complete reporting of the data, including the ways that the study was approved and relevant details of the sample. Authors are required to report either the ARRIVE or PREPARE guidelines for animal research in the submission.

We recommend that authors replace, reduce, and refine animal research as promoted by the N3RS.

Data sharing policy

If any original data analysis results are provided within the submitted work, such results have to be entirely reproducible by reviewers, including accessing the data.

Submissions are, by definition, contained within one or multiple repositories, which we require authors to archive before we accept the submission. Any data contained within the software is made available accordingly.

Human participants research policy

In the exceptional case a JOSS submission contains original data on human participants, the study must have been performed in accordance with the Declaration of Helsinki. The manuscript must contain all relevant information regarding ethics approval, including but not limited to the responsible ethics committee and reference number. This also applies to ethics exemptions. Informed consent must be collected from all human research participants, and authors are required to state whether this occurred in the manuscript.

AI usage policy

The Journal of Open Source Software permits the use of generative AI in submissions with mandatory disclosure and human oversight requirements.

Author AI usage

The use of generative AI is permitted for most aspects of a JOSS submission (e.g., software creation and review, generating documentation, assisting with paper authoring).

Required disclosure: All submissions must include an AI usage disclosure section in the paper that clearly states whether generative AI tools were used. This disclosure must include:

  1. Tool identification: Specify which AI systems and versions were employed, noting exactly where they were applied (code, documentation, manuscript sections).

  2. Scope of assistance: Describe the nature of support provided—examples include code generation, refactoring assistance, testing scaffolding, editorial review, or manuscript drafting.

  3. Human verification confirmation: Authors must affirm that human team members thoroughly reviewed, modified, and validated all AI-generated content while making primary architectural and design decisions.

Prohibited AI interactions: Conversational use of AI between authors and editors/reviewers is restricted, except for translation support.

Author accountability: Submitting authors bear complete responsibility for accuracy, originality, licensing compliance, and ethical/legal standards of submitted materials. Incomplete or inaccurate AI disclosures constitute ethical violations with consequences ranging from desk rejection to institutional notification.

Reviewer AI usage

Reviewers may employ AI tools for non-substantive tasks (grammar checking, code contextualization) requiring brief disclosure at review conclusion.

Critical constraint: All evaluative determinations—scoring, recommendations, originality assessments, and compliance judgments—must reflect human reviewer judgment exclusively.