Facilitating scientific discourse fairly and responsibly is a fundamental function of scientific publishing. Science is a highly communal endeavor; on one hand, interactions between researchers can play unique and powerful roles in driving progress. Consider, for example, that Bayes’ theorem was only submitted to the Royal Society’s journal Philosophical Transactions by one of his friends in 1763, two years after Bayes’ death1. But at the same time, just as Bayes’ theorem carries the name of its discoverer, there is also importance given to ownership of intellectual contributions. As such, promoting transparency in scientific publishing is critical to supporting the community, from individuals to the collective.

Credit: Vasko/E+/getty

The broader scientific community has devised various approaches to transparency in publishing for hundreds of years, and likely longer. In 1610, for example, while observing the phases of Venus that would help overturn the geocentric view of our Solar System, Galileo Galilei sent word of his findings in the form of an anagram to Johannes Kepler. It is believed that this was a way to avoid being preempted by other observers while he validated his predictions for the phases of Venus: “Haec immatura a me iam frustra leguntur o.y.”. Or, in English: “These, premature from me, are at present deceptively gathered together”2.

While the scientific community today no longer has to rely on puzzles in Latin to protect its findings, the question of how to optimally support transparency in scientific publishing while protecting the interests of its individual contributors remains an open problem filled with subtleties. Naturally, then, one of the primary goals of Nature Chemical Engineering is to develop and support tools and practices to help our authors, reviewers and readers navigate these challenges. In this Editorial, we cover some of the new and established mechanisms through which we are promoting transparency throughout our publishing process.

First, transparent peer review is now going to be available on all the Nature Research journals. As of July 2025, our transparent peer review, which authors can choose to opt into upon seeing the referee reports at any resubmission stage, allows readers to view both the referees’ comments and the authors’ response to those comments. Transparent peer review was first introduced to the portfolio in Nature Communications in 2016, followed by Nature and several of the Nature Research journals in 2020. Much like listening to questions and discussions at conferences, we believe that transparency in peer review gives readers a glimpse into the valuable discourse that has helped shape the final published manuscript.

Other key aspects of transparency are the availability of custom code and data sharing. We already require that authors make custom code available via a public or private repository to referees when the work’s contributions hinge substantially on that custom code. As of early 2025, we now offer authors support to upload their code to a service called Code Ocean, wherein the authors’ code ‘capsule’, which contains a pre-installed environment for code testing, is supplied to the referees3. One important transparency distinction between these two options is that Code Ocean capsules become public with their associated papers, whereas repositories can be kept private on a case-by-case basis (for example, due to third-party control), with any restrictions to access clearly described in a code availability statement. And while our journal policy on data sharing has remained unchanged since our launch, wherein each published primary research article includes a data availability statement along with ‘source data’ for manuscript figures, we also highlight a recent Nature Materials Editorial that describes our data sharing policies in detail4.

Finally, as large language models (LLMs) continue to gain widespread traction as multi-purpose assistants, we want to highlight several aspects of our current policy related to transparency in the use of artificial intelligence (AI) tools to prepare submission materials. First, we do not allow the inclusion of generative AI images in our publications (such as manuscript figures and cover art). And second, we do not attribute authorship to AI tools, as we do not consider these tools to meet our criteria for authorship, particularly in terms of accountability. Moreover, although these tools can be used for wording and formatting changes to texts without the need for declaration, AI assistants should not be utilized for generative editorial work and autonomous text creation. LLMs and other AI tools can, however, be used in other aspects of research so long as their use is appropriately described in the Methods section of the submission. Given the rapid evolution of this space, this policy may evolve with time; we will advise our community of these updates if or when they occur.

One of our most important responsibilities as a scientific journal is to promote transparency in scientific publishing in a way that is both useful to our community and in the best interest of our authors’ intellectual contributions. And while we do not discourage colleagues from sharing word puzzles with each other, we hope that our community will find the initiatives outlined in this Editorial more effective tools to promote transparency in disseminating their findings than Galileo’s coded message to Kepler, whose solution, for the interested reader, was “Cynthiae figuras aemulatur mater amorum”, or roughly translated, “Venus imitates the shape of the moon”.