Science is a force for good. The federal government funds scientific research because the returns belong to all Americans. But that investment only pays off when results reach the people and institutions that can act on them. Research locked away from the public, from entrepreneurs, from clinicians, from other scientists, is research whose value has been cut short.

Communicating results is part and parcel of doing scientific research

Scientific research only gains in value when the results are shared as widely as possible, so they can be interrogated, improved, and built upon. This is not a secondary consideration. It is intrinsic to how science works. A finding that cannot be accessed, examined, or built upon is a finding that cannot fully deliver on the public investment behind it.

Fixing the scientific communication system will require honest choices. The commercial publishing industry has played a real role in organizing and disseminating research, and some publishers are genuinely adapting to a more open future. But the structural incentives of the current system are working against the public investment this work is here to protect, and Congress has an opportunity to change them. That opportunity is immediate: the House Science Subcommittee held a hearing this April on exactly these questions, and members on both sides of the aisle left with more concerns than answers about who controls the scientific record, and whose interests that control serves.

The current system is not delivering 

The hearing documented what happens when scientific communication is not working: paper mills, fraudulent research, a reproducibility crisis, and findings that cannot be verified or built upon. These are not peripheral problems. They are what happens when a scientific communication system is optimized for something other than the open exchange of knowledge.

Peer review plays a critical role in helping the scientific community and the public understand how to interpret and contextualize research results. But peer review as it is currently being deployed by publishers, in a limited, closed, slow, proprietary, and uncredited way, is not working. We currently find ourselves waiting years for peer-reviewed articles to be shared, only to discover they are often flawed, and we cannot examine the underlying review. Yet these articles are then held up as an arbitrary measure of quality, as if they are somehow sacrosanct.

What the hearing made clear is that peer review’s failures are not accidental. They are structural. Publishers control the process — who reviews, how reviews are conducted, whether they are visible — and they do so with no particular accountability to the scientific community or the public. As Rep. Zoe Lofgren observed, the peer reviewers are other scientists who donate their time.  The question of who benefits from controlling that process, and who gets to set its rules, deserves far more scrutiny than it currently receives. The most visible expression of that control is what publishers call the ‘version of record’ — a concept that exists not to serve science, but to anchor the commercial publishing business model.

The ‘version of record’ is a term created by the publishing industry to serve its own profit-making interests. It treats a single formatted snapshot as the definitive form of a scientific finding, when science itself is a living, self-correcting process. – Heather Joseph, SPARC Executive Director

A system built around this singular “version of record” is a system that freezes science in commercial interests over the interests of the public that paid for the research.

What a better system looks like

What science actually needs is a rich record of versions: results communicated early and often, shared as widely as possible so they can be interrogated, improved, and built upon, with the understanding that each represents what we know right now. Supporting ongoing research to feed new information into the mix is what science is all about.

Peer review should be part of that open process. Peer review should happen earlier and more often; it should be conducted openly, where possible and appropriate; and it should be conducted non-anonymously, so researchers can be attributed and credited for the work they contribute. Platforms like eLife, Review Commons, and Peer Community In already demonstrate that this is possible. arXiv, bioRxiv, and medRxiv have shown at scale that rapid, open dissemination followed by community review produces high-integrity, widely useful science.

At the hearing, Kate Travis noted that open data, open datasets, and open peer review are among the most practical tools available to improve reproducibility as working practices the scientific community is already adopting. We can build a research communication environment that prioritizes the early and open sharing of results on platforms where the scientific community can organize and conduct peer review in ways that are appropriate to their disciplines. This is not a theoretical vision. It is already happening, and it works.

What Congress should do

These problems are connected, and addressing them requires treating them as a system. The April 15 hearing raised important concerns about business models, AI-generated slop, data security, and foreign access to federally funded research — all real issues. But the thread running through all of them is a scientific communication system whose structure is set by commercial publishers with interests that don’t always align with the public good. We offer the following framework:

• Open by default. Congress should codify and strengthen the 2022 OSTP guidance requiring immediate public access to federally funded research, making openness the baseline expectation, not a negotiated afterthought.
• Invest in community-owned infrastructure. arXiv, bioRxiv, PubMed Central, and a growing ecosystem of nonprofit and diamond open access journals demonstrate that rigorous scientific communication can be organized outside the commercial market. Congress should invest in this infrastructure directly through NSF, NIH, and IMLS.
• Reform researcher assessment. The pressure to publish in high-prestige commercial venues sustains many of the dynamics the Subcommittee identified. NIH and NSF could require that grant applicants be evaluated on the quality and openness of their outputs, realigning incentives toward scientific quality rather than journal prestige.
• Mandate transparency. The reproducibility crisis is at its core a transparency crisis. Mandating open data, methods, and code for federally funded research — and building the repositories to support compliance — would do more for research integrity than almost any other intervention.

Science deserves a communication system built for science

Scientific communication should ensure that researchers are communicating regularly with one another, with funders, and with the public. The federal government funds science because its returns belong to everyone. The communication system should reflect that purpose.

The April 15 hearing showed a real bipartisan appetite for fixing this system. What it also showed is that the publishing industry’s testimony focused, understandably, on preserving the industry’s role. The question for Congress is whether that role, as currently structured, is serving American science or constraining it. American science has the talent, the institutions, and the infrastructure to lead the world in open, transparent, rigorous research. What it needs is a policy framework that makes openness the default and ensures publicly funded knowledge reaches the public that paid for it.