When the Public Library of Science, a non-profit organization based in San -Francisco, California, and other publishers popularized article-processing charges (APCs) in the mid-2000s, scholarly publishing in Latin America was already embracing open access (OA) using a different model: instead of charging authors, academic institutions published journals edited by faculty members. The approach is a type of ‘diamond OA’, which works without fees for readers or authors.

Over the same time period, APCs have become ubiquitous in the global north, embraced by for-profit journals and encouraged by many leading European and US funders. The vibrant publishing ecosystem in Latin America (and elsewhere in the global south) will not be left unscathed.

I know this ecosystem is vibrant and diverse because I’ve spent 15 years working at the Public Knowledge Project (PKP) — based at Simon Fraser University in Vancouver, Canada — and I am now co-scientific director of this initiative to make research publicly available. I have met hundreds of journal editors who work hard, often in challenging conditions, to bring the knowledge discovered by their communities to the rest of the world. An incredibly diverse set of journals now uses Open Journal Systems, free software developed by PKP, to manage, publish and index their work (S. Khanna et al. Preprint at SciELO preprints https://doi.org/jgbz; 2022). My and my colleagues’ work shows that many non-academics frequently access this content. Many titles focus on locally relevant issues, such as rural development, local histories and Indigenous cultures. There are, of course, journals of broad global interest as well.

Yet many Latin American scholars still seek to publish in well-funded journals in Europe and North America, which are seen as more prestigious and get more international attention. And many research-evaluation systems favour journals indexed in databases such as Scopus and Web of Science, which contain a tiny fraction of the journals found in the global south. As more journals charge APCs, more Latin American institutions are pressured to pay them.

In Colombia alone, APC payments are estimated to have grown by 18-fold since 2019. The amount is expected to increase after some five dozen institutions signed Latin America’s first ‘transformative agreement’ (a contract to pay APCs to subscription-based journals that are changing business models) late last year. At least 120 journals in Latin America have begun charging APCs in the past 5 years, although this model inherently links publication to authors’ (or their funders’ or institutions’) ability to pay.

APC waivers cannot solve this problem. They require authors to proactively ask for charity (even as Latin American institutions shoulder diamond OA costs), and rarely apply to countries in the region, most of which are above income thresholds set by publishers. As institutions pay more APCs, they will feel pressure to forgo investments in the Latin American diamond OA ecosystem.

There is some movement towards diamond OA in the rest of the world. In September, I participated in the first Diamond Open Access Conference, hosted online and in Zadar, Croatia. This year, the European Commission has awarded funding to two initiatives to support institutional OA publishing. In Canada, national funders continue to support the PKP and its partnership with the dissemination platform Érudit to build open, non-commercial infrastructure for sharing research. In August, Redalyc, an OA initiative at the Autonomous University of Mexico State in Toluca, was awarded a grant from Arcadia, a charitable fund in London. Signatories of the Action Plan for Diamond Open Access (see go.nature.com/3cghq) include more than 130 research-supporting institutions across the world. This support is welcome, but more is needed.

Of course, the model is not perfect. OA funding mechanisms are needed that protect the diversity that exists in the system, lower the burdens of production and raise the quality of journals. Reliance on volunteers and institutional support (such as access to servers, software and IT staff) make some diamond OA journals precarious. However, in my view, this model is much less problematic than one in which journals gain more revenue for publishing more papers and that relies on publishing papers from fields and communities that are well resourced.

APCs beget APCs. The more funds that are available to pay them, and the more researchers who have the ability to do so, the more journals will feel pressured to charge them.

If governments, funders and institutions — including those in Latin America — do not want to be responsible for dismantling this diverse and global scholarly OA ecosystem, they should stop supporting APCs altogether. Funds that are allocated to APCs should be invested in shared infrastructure, tools and services that can support multiple journals simultaneously.

Such an embrace of diamond OA could lead to virtuous cycles in which journals can lower operating costs, raise their quality and elevate their place in research assessment.

This future might not seem possible to those ensconced in an APC-world dominated by large, commercial publishers. I have worked alongside editors, scholarly publishing experts and OA advocates long enough to know that it is — but only if APCs stop before the two models collide and we get thrown off the rails.