Building a Knowledge Commons

In “Constructing a New Knowledge Infrastructure” (Issues, Spring 2026), Shannon Dosemagen and Gwen Ottinger propose an ambitious, decentralized platform for crowdsourcing environmental pollution data to empower local communities and advance evidence-informed policymaking. The idea of building such a data commons is resonant and timely as the Trump administration dismantles much of the knowledge infrastructure that communities, scholars, and policymakers have long relied on and, in some cases, abdicates entirely its responsibility to collect, synthesize, and communicate science to the public.

History tells us that an effective tool against such government actions is to build parallel institutions that fulfill the functions abandoned by the government. A knowledge commons could be a key component of this strategy, providing reliable environmental data to empower people to act and counter any disinformation from political forces. The idea is also valuable because environmental justice communities often have limited access to the data and analyses needed for effective policy advocacy, and recent state-level attempts to ban community science may further restrict such data availability. Sharing across communities is crucial for public health, especially as the administration’s rollbacks of environmental regulations are likely to worsen the pollution exposure inequities and cumulative impacts that already burden Black and Brown communities disproportionately.

Beyond ensuring ready access, it’s also important to share how data are used in policy contexts; communities often need an understanding of how to leverage that scientific information in regulatory processes, such as permitting, zoning, and emissions control decisions. I have seen the complexity and nuance of environmental policy weaponized against communities and used to justify inaction by government; it is this interface between science and policy decisions where knowledge-sharing across communities could improve their outcomes.

Several challenges, not addressed directly in the article, must be navigated for such a project to succeed, including the quality control, interoperability, and proper interpretation of information. As the authors note, a project is more durable if it is decentralized without the need for full-time administrative support, but this leaves the challenge of who is responsible and how to ensure data are of adequate quality and standardized across localities. The risk of misuse and misinterpretation could undermine the cause, in a world where community leaders’ data and lived experience are already often not taken seriously by decisionmakers.

Let’s pilot the idea. Communities nationwide are concerned about the local environmental consequences of data centers; thus, this salient and timely issue provides an opportunity to test participation levels and utility of such a model. The nonprofit EJ35, for instance, recently built a screening tool that maps proposed and existing data centers overlayed with public air quality information to provide residents, organizers, journalists, and policymakers with credible transparent analysis. With tools such as these, communities can share data and policy advocacy information, and the functionality of such a knowledge commons could be tested. The current moment makes clear we need to reimagine science and knowledge infrastructures at large: as the current administration breaks things, let’s build something better.

Shannon Dosemagen and Gwen Ottinger correctly observe that creating effective governance structures is the aspect of the environmental knowledge commons that will require the most care, experimentation, patience, and goodwill. They are right. The cases they examine also point toward a governance layer their essay helps bring into view.

The Public Lab nonprofit they cite shuttered in 2023 when it could no longer raise funds to maintain the resources the community had amassed. That failure is a governance story about what happens when no named function exists to carry the value of a knowledge commons into decisions determining its survival. The knowledge existed. Yet, because grassroots contributors are routinely consumed by the urgency of acute local crises, it had no route into the room where the decision about its future was made.

The same failure mode appeared at a larger scale in March 2026, when the US Department of Agriculture announced the closure of 57 Forest Service research facilities. Targeted in that announcement, the Hubbard Brook Experimental Forest in New Hampshire holds over 60,000 barcoded samples spanning seven decades. Nobody collecting water samples there in 1963 was thinking about PFAS contamination, the “forever chemicals,” but those physical samples are why scientists can establish preindustrial baselines that would otherwise be impossible to reconstruct. When the closure decision was made, that value had no institutional representative. No governance requirement compelled decisionmakers to account for decades of archival option value. The archive existed, and the evidence was real. It had no route into the room.

These cases share the absence of a named translational function: someone responsible for carrying what the commons produces into the decisions that shape its survival. This is a different function from stewardship of the commons itself. It ensures the commons is represented when administrative decisions are made, and that the experience of those decisions returns to the knowledge system.

Dosemagen and Ottinger argue funders must restructure incentives to support long-term stewardship. That is necessary. But a grant condition that funds stewardship infrastructure without naming who is responsible for the translation function risks reproducing the same vulnerability: The commons is maintained, but its value is absent when it matters.

The systems that deliberately address this problem make continuity structural rather than cultural. The Global Climate Observing System’s Reference Upper-Air Network and the European Research Infrastructure Consortium’s Integrated Carbon Observation System use named stewardship obligations, formal site responsibilities, and institutional commitments that survive personnel transitions. The funder-mandate the authors call for is the lever for requiring the same discipline at the translational layer: Explicitly name who is responsible for carrying evidence into decisions, document how that function survives transitions, and evaluate it as a governance obligation rather than a checkbox.

The environmental knowledge commons will not reach its potential through data infrastructure or collaborative governance alone unless the translation function is explicitly named. Dosemagen and Ottinger make the infrastructure problem visible. The next step is naming the governance layer beneath it.

Shannon Dosemagen and Gwen Ottinger diagnose the environmental knowledge commons problem with refreshing clarity and honesty. Their insistence that hostile federal administrations are not the primary obstacle—and that short-term, place-bound funding undermines both the infrastructure commons require and our imagination about alternatives—names challenges I know well. Farm Hack, which I cofounded, and Public Lab, one of their case studies, emerged and cross-pollinated in the same period, convinced that communities could build and govern their own knowledge. Although both started as nonprofits, Farm Hack has arrived where the authors do: Commons governance and decentralization is not an add-on but part of the infrastructure itself—technical, social, legal, and financial. I would extend it in two directions.

The first concerns reciprocity. When modes of exchange, attribution, and consent are built in, contribution becomes generative rather than extractive: A farmer, rancher, or land steward who sees knowledge attributed, consent honored, and value returned gives generously; open exchange compounds what each brings. So open licensing, though necessary, is insufficient—it governs access, not purpose or utility. It does not answer the questions that matter most: Will value return to those who generated it? Who decides when it is combined with datasets they never anticipated? Will they retain access? A consent layer that honors these conditions keeps trust and sovereignty at the center, turning a repository into a living commons.

The second concerns what makes the decentralized model reproducible. The contrast between Public Lab and the Reagent Collaboration Network, or Reclone, that the authors cite is important: a handful of committed stewards maintaining shared resources that many communities draw on, with no single owner. Reclone’s founders also brought established labs and networks; most farming communities start without that baseline. The reproducible lesson is the structure, not the starting resources; two things can supply what these communities often lack—a governance vehicle and baseline technical capacity.

On governance: As the authors observe, the nonprofit model itself limited Public Lab; funders who award to a single lead organization push toward ownership and control, not collective stewardship. We need purpose-built community governance models the nonprofit form was never meant to provide. I propose we define a category of Commons Governance Organizations defined by what they are, not what they are not: chartered to hold shared knowledge and assets in trust.

On capacity: We are piloting one approach with the Farm Hack Box—a self-hosted, offline-first community node on a $400 Raspberry Pi computer drawing 27 watts with 1 terabyte of storage, providing access to a curated global knowledge commons, preserving data sovereignty, locked to no vendor or service. It is a great evener, giving a rural cooperative or seed library the federated data-sharing capability a university lab takes for granted; it enables resilience with neighbors and persists as a community archive even if an agency abandons a program. More tools than ever can close this gap; the Box is one reflection of the potential for community-governed and -held knowledge utilities.

We offer these as contributions to the conversation the authors have opened, eager to help develop commons-enabling infrastructure in the field.