PLOS’s Own Authorship Policy Does Not Mention AI at All

Most of PLOS’s authorship page is conventional and worth stating plainly because it is useful: authors must meet three criteria — substantial contribution to the work (conception, design, data, software or drafting), approval of the submitted and any substantially modified version, and personal accountability including helping investigate accuracy or integrity questions. Every contribution, from everyone meeting these, must be reported using the CRediT taxonomy, a standard list of contribution roles. Group authorship follows ICMJE guidance, and an individual contributor still has to be designated the corresponding author even when a consortium name appears in the byline.

What is missing is the point of this brief. We read the authorship page in full, and separately read PLOS’s best-practices-in-reporting page, looking specifically for any rule on AI authorship, AI disclosure, or AI in figures. Neither page uses the words “AI”, “artificial intelligence”, “chatbot”, “generative” or “large language model” anywhere in the body text.

The one disclosure requirement the authorship page does carry, and the nearest thing to a transparency rule on it, concerns a different kind of assistance entirely: “the involvement of any professional medical writer in the publication process must be declared”.

It would be easy to read a publisher’s silence on AI as permission, and that would be a mistake: silence is not the same as a stated absence of a rule, and a journal editor handling your submission may apply an unwritten expectation, a cross-reference to ICMJE (which PLOS says its authorship criteria are based on, and which does have an AI section as of its January 2026 update — see our companion brief), or their own judgement, none of which this brief can predict for you.

The gap is also worth naming because of what it implies about the page itself: a policy document that has not been revisited to address AI since generative tools became common newsroom-and-lab equipment is, at minimum, a document whose silence you should not treat as settled guidance.

What PLOS does require — the CRediT taxonomy and the three authorship criteria — still functions as a baseline a reader can apply directly: if a contribution does not meet the three criteria, it is acknowledged rather than authored, whatever produced it.

How it reaches a specific journal. The authorship criteria and CRediT requirement apply across the PLOS portfolio read here; PLOS Medicine is the named exception with its own authorship statement, which this brief has not separately verified and does not summarise.

Where your target journal’s own instructions do mention AI — some PLOS journals may have added guidance since this brief was checked — that specific page governs; the absence documented here is about the general authorship and best-practices pages read on the date shown, not a claim that no PLOS journal anywhere addresses AI.