3 Critical Assumptions in FAIR and CARE Data Principles for Interoperability and Epistemic Plurality

Introduction

The FAIR data principles—findability, accessibility, interoperability, and reusability—have become a global benchmark for data accessibility standards in science. Developed to make research outputs more discoverable and reusable, they align closely with the goals of open science and reproducibility. In parallel, the CARE data principles—collective benefit, authority to control, responsibility, and ethics—emerged from Indigenous data governance to safeguard cultural data sovereignty and ensure ethical data sharing in ways that respect the rights and values of the communities from which the data originate.

Nowadays, both FAIR and CARE are central to debates on biodiversity monitoring, cultural heritage preservation, and large-scale data integration. Advances in interoperability in biodiversity data now enable connections between genomic repositories, environmental time-series, and acoustic ecosystem monitoring platforms. However, such technical achievements also raise concerns: which standards define accessibility, how epistemic diversity is represented, and whether universal schemas risk erasing essential cultural and contextual meaning.

At the heart of this conversation lies a persistent tension between interoperability—the ability of diverse systems to exchange and use information—and epistemic plurality, the recognition that different knowledge systems hold equally valid but sometimes incompatible ways of knowing. Addressing this tension requires more than technical solutions; it demands a commitment to data ethics in science as a transversal framework that connects governance, infrastructure design, and cultural responsibility.

A. Conceptual Foundations

1. Origins and Purpose of FAIR Data Principles

The FAIR data principles were formally articulated in 2016 through a publication in Scientific Data, emerging from an international effort to address systemic problems in data discovery, interoperability, and reuse. Designed to make datasets more findable, accessible, interoperable, and reusable, the principles sought to create a shared baseline for data accessibility standards in an increasingly interconnected research ecosystem. While often described in technical terms, FAIR also reflects specific epistemic commitments: that openness and standardization inherently benefit science, and that efficiency in data exchange is a universal goal.

In biodiversity and genomics, FAIR underpins major repositories such as the Global Biodiversity Information Facility (GBIF) and GenBank. These platforms exemplify large-scale implementations of interoperability, allowing researchers to cross-reference species occurrence records, molecular sequences, and ecological metadata at unprecedented scales. Yet their prominence raises important questions: Who decides which data qualify for inclusion? How are culturally sensitive species names or geospatial coordinates handled? To what extent do global repositories account for local governance rules or community consent? And what mechanisms exist for redress if data are used in ways that violate the original intent of contributors?

FAIR’s focus on machine-actionable integration assumes that standardization can occur without erasing context. However, for certain datasets—particularly those involving Indigenous data governance or endangered species—disclosure can pose ethical risks, from biopiracy to cultural appropriation. In such cases, rigid adherence to FAIR’s technical mandates may conflict with responsible data management practices that emphasize relational accountability over universal openness. This has led to growing interest in integrating FAIR with the CARE data principles, not as an afterthought, but as a co-equal framework that interrogates the very assumptions about who benefits from “open” data.

2. Origins and Purpose of CARE Data Principles

The CARE data principles—collective benefit, authority to control, responsibility, and ethics—were introduced by the Global Indigenous Data Alliance (GIDA) as a direct response to gaps left by the FAIR data principles. Where FAIR privileges machine-actionability and data accessibility standards, CARE elevates relational accountability, Indigenous data governance, and cultural data sovereignty. Yet the move from values to operations is complicated: What precisely counts as “collective benefit”? Who has the legitimate authority to control in heterogeneous communities? And how can ethical data sharing be sustained when national mandates or funder policies demand openness that communities may not consent to?

Historically, CARE emerges from decades of advocacy confronting extractive research practices—cases where ecological, biomedical, or cultural datasets were published without free, prior, and informed consent; where location data jeopardized sacred sites or endangered species; and where knowledge shared for stewardship was later commercialized elsewhere. CARE reframes openness: instead of assuming that more visibility is intrinsically good, it asks whether data flows produce equitable outcomes and whether consent is ongoing rather than a one-time checkbox. But this reframing invites practical dilemmas for responsible data management: if consent is dynamic, can withdrawal be meaningfully honored once data are mirrored, indexed, or used to train machine learning models? If not, what compensatory duties—like benefit-sharing or community veto power—become ethically non-negotiable?

Implementing CARE within large infrastructures—biodiversity portals, genomic repositories, or cultural archives—requires rethinking architecture. How should repositories encode governance constraints in machine-readable ways so that access rules travel with the data? Are current metadata schemas capable of representing community protocols, embargo periods, or conditional use clauses, or do we need new policy-aware ontologies? If interoperability in biodiversity data is achieved by flattening local meaning for cross-system compatibility, do we risk undermining epistemic plurality? Conversely, if we fragment standards to reflect local nuance, do we erode the very interoperability that enables discovery, reuse, and global collaboration?

CARE also surfaces internal questions often overlooked in policy debates. Communities are not monoliths: who speaks for whom, and through what processes? What happens when youth, elders, and diaspora disagree on the terms of reuse? If authority rests with a designated council today, how are future governance transitions handled so that long-term datasets do not become stranded or overexposed? And in multi-jurisdictional contexts—where a species range, river basin, or oral tradition crosses borders—how do repositories reconcile conflicting legal regimes and overlapping community claims to stewardship?

There is a further operational risk of “CARE-washing”: adopting the language of CARE without changing incentives, data pipelines, or access controls. Checklists and static consent forms are insufficient when data move at API speed. Real alignment demands auditable processes: explicit community review stages; versioned consent and licensing; the use of granular access tiers (open, group-restricted, time-bound, and sensitive-use); and traceable decision logs that document why data were released, restricted, or withdrawn. Without such mechanisms, CARE remains aspirational, offering powerful rhetoric but limited leverage over practice.

In technical terms, enforcing CARE at scale likely requires policy-aware infrastructure: access brokers that evaluate user intent and affiliation; machine-readable governance metadata that travel with datasets; dynamic licenses (e.g., Traditional Knowledge labels or community-specific terms) that can be updated; and audit trails that alert custodians to secondary uses. Ethically, it implies redistributing control—embedding community veto points into workflow, honoring moratoria on sensitive locations, and formalizing benefit-sharing beyond acknowledgments. Legally, it raises questions about the enforceability of community protocols against downstream users in other jurisdictions, particularly when data are copied or integrated into derivative products like trained models or aggregated indexes.

• Collective Benefit: Who defines it, how is it measured over time, and what remedies exist if benefits do not materialize?
• Authority to Control: What constitutes legitimate representation, and how are internal disagreements adjudicated?
• Responsibility: Which actors (repositories, funders, journals, developers) bear duties to prevent misuse and to remediate harm?
• Ethics: Can machine-readable policies capture situated ethics, or do we risk reducing relational duties to formal toggles?
• Interoperability vs. Plurality: How far can we push standardization before it erases context essential to cultural data sovereignty?
• Lifecycle Consent: How are consent, embargo, and withdrawal operationalized once data are mirrored or used for model training?

Framed this way, CARE data principles are not merely a supplement to FAIR; they are a structural critique of how infrastructures define value, authority, and accountability. They push data ecosystems to demonstrate—not just declare—respect for Indigenous data governance and to reimagine ethical data sharing as an ongoing relationship, not a one-time transaction. The challenge for today and beyond is to transform these questions into enforceable design patterns and governance routines that scale without sacrificing the communities whose knowledge makes the data possible in the first place.

3. Defining Epistemic Plurality

Epistemic plurality refers to the coexistence of multiple, equally legitimate ways of generating, validating, and interpreting knowledge. In scientific contexts, it challenges the assumption that a single epistemic framework—often rooted in Western scientific traditions—can fully capture the complexity of the world. While interoperability in biodiversity data aims to connect datasets through common standards, epistemic plurality reminds us that what counts as “data,” “evidence,” or “valid knowledge” is not universal, but culturally and contextually determined.

This concept is particularly relevant in biodiversity and environmental sciences, where knowledge may be encoded in peer-reviewed taxonomic checklists, but also in oral histories, seasonal calendars, or ritual practices that defy digitization. How can a database architecture honor these different epistemologies without reducing them to fields and codes in a metadata schema? Does translating Indigenous ecological knowledge into standardized vocabularies inevitably strip it of meaning, or can careful co-design preserve nuance while enabling discovery? And when conflicts arise between knowledge systems—such as differing criteria for species identification—who decides which perspective prevails in the integrated record?

The tension between epistemic plurality and responsible data management is not merely technical. It is also about governance and legitimacy: whose voices are included in the decision-making process about data standards, and whose are systematically excluded? For example, biodiversity platforms might adopt FAIR-aligned data models that streamline analysis but inadvertently exclude phenological indicators known only to local fishers or pastoralists. Is this a trade-off we accept for the sake of interoperability, or a sign that our interoperability goals need to be redefined?

In policy circles, epistemic plurality is sometimes treated as an abstract principle—important in theory, but difficult to operationalize. Yet ignoring it can have concrete consequences: conservation strategies that fail because they overlook local indicators of ecosystem change; genomic surveillance systems that miss culturally relevant markers; or benefit-sharing agreements that falter because they do not account for differing conceptions of ownership. The integration of CARE data principles alongside FAIR is one attempt to address this, but even here the question remains: can a single infrastructure truly accommodate multiple epistemologies, or will one inevitably dominate?

• Definition: How do we define “knowledge” in ways that respect diverse epistemologies without collapsing them into a single model?
• Representation: Can data schemas be designed to carry context and cultural meaning alongside technical attributes?
• Conflict Resolution: Who mediates when epistemologies offer contradictory conclusions about the same phenomenon?
• Governance: What mechanisms ensure that marginalized epistemologies are not tokenized but meaningfully integrated?
• Ethics: How do we prevent epistemic extraction—borrowing concepts without honoring their origins or obligations?

Ultimately, epistemic plurality demands that interoperability be reframed not as a purely technical achievement, but as a relational and ethical practice. The challenge for today and beyond is to design infrastructures, policies, and collaborative norms that allow knowledge systems to meet without forcing them into premature consensus—preserving the very diversity that makes global data ecosystems resilient and just.

B. Hidden Assumptions in FAIR and CARE

1. Universalism in Data Models

One of the least examined premises of both FAIR data principles and certain implementations of the CARE data principles is the belief in universal applicability: the idea that a single data model, once standardized, can serve the needs of all stakeholders equally. This notion of universalism is attractive because it promises efficiency, scalability, and predictability in responsible data management. However, it also embeds a series of epistemic and political assumptions about what counts as valid knowledge, what metadata are essential, and how data should be organized for reuse.

In biodiversity informatics, for example, universalism manifests in taxonomic backbones, geospatial coordinate systems, and controlled vocabularies used by platforms like GBIF. These structures make cross-dataset queries possible, yet they implicitly privilege certain ontologies over others. What happens when an Indigenous classification system for a plant does not map neatly onto Linnaean taxonomy? Does the system accommodate multiple classifications, or does it overwrite local knowledge in the name of interoperability? The same questions apply in genomics: when GenBank assigns identifiers, whose naming conventions are prioritized, and which metadata are stripped away as “non-essential” for the global user base?

The universalist impulse also assumes that data models can remain neutral, overlooking the fact that their very design choices—how fields are named, which data types are supported, what categories are mandatory—reflect the priorities and blind spots of their creators. A schema optimized for climate modeling might downplay culturally important seasonal markers; a structure built for museum collections might fail to represent community restrictions on sacred object images. By asserting universality, these models risk encoding biases into the technical infrastructure, making them harder to detect and challenge over time.

Critics argue that instead of aiming for one-size-fits-all universality, infrastructures should pursue “bounded interoperability,” where systems can connect but also maintain space for local variation, alternative ontologies, and culturally specific metadata. Yet this raises its own questions: How much heterogeneity can an infrastructure tolerate before interoperability breaks down? Who decides which differences are worth preserving and which are collapsed into a common standard? And what governance structures are in place to revisit these decisions as knowledge systems evolve?

• Scope: Is the goal to connect all data under one model, or to allow coexistence of multiple models linked by translation layers?
• Authority: Who determines the “core” metadata fields that define compliance with FAIR or CARE?
• Adaptability: Can universal models evolve quickly enough to incorporate emerging epistemologies and contexts?
• Risk: How do we prevent the erasure of culturally or locally significant data elements during standardization?

Universalism in data models, while efficient on paper, often conceals the power dynamics that shape whose data practices are made legible and whose are marginalized. The challenge for today and beyond is not simply to expand the reach of a single model, but to build infrastructures capable of negotiating boundaries—allowing for shared technical ground without flattening the diversity of ways in which the world is known and represented.

2. Accessibility vs. Control

The tension between accessibility and control sits at the heart of the friction between the FAIR data principles and the CARE data principles. FAIR frames accessibility as a positive norm—data should be as open as possible, with restrictions applied only when necessary. CARE reframes this assumption, arguing that in many contexts, cultural data sovereignty and responsible data management require conditional, negotiated access based on community-defined rules. Both positions claim to serve the public interest, but they diverge sharply on who decides what “public” means and what forms of access are acceptable.

In biodiversity repositories, FAIR-aligned accessibility means that species occurrence data, once digitized, can be queried, downloaded, and reused globally. Yet the publication of precise geospatial coordinates for certain plant or animal species can facilitate exploitation—illegal collection, poaching, or bioprospecting. CARE asks whether “open” access to such information truly serves collective benefit, or whether it undermines stewardship by enabling harm. Should a community be able to embargo sensitive ecological data indefinitely, even if doing so limits scientific analysis? And if so, how is that decision recognized and enforced in infrastructures optimized for instant retrieval?

The control side of the equation also raises operational dilemmas. Some Indigenous communities have established digital repositories with multi-tiered access: open data for non-sensitive materials, restricted access for culturally specific content, and closed archives for sacred or ceremonial knowledge. Integrating such tiered models into FAIR-compliant systems is technically possible, but often politically fraught. Who verifies user credentials for restricted datasets? How are access decisions logged and contested? Can machine-readable licenses reflect the fluidity of consent, or are they too rigid to accommodate shifting community priorities over time?

The accessibility-control debate exposes a broader question about ethical data sharing: is openness a default setting, or a negotiated outcome? FAIR implicitly assumes the former; CARE insists on the latter. Neither position alone resolves the tension, and attempts to merge them—such as hybrid metadata schemas or “consent-aware” APIs—face both technical and governance hurdles. Without sustained dialogue between infrastructure developers, policymakers, and rights-holding communities, there is a risk of settling for surface-level integration that preserves the rhetoric of both frameworks while diluting the protective force of CARE.

• Risk Assessment: Who determines when the benefits of openness outweigh the potential harms of disclosure?
• Enforcement: What mechanisms exist to enforce community-imposed access restrictions across interoperable systems?
• Flexibility: How can access controls adapt to changes in community governance, environmental threats, or new scientific uses?
• Accountability: Who is responsible when data are accessed or used in ways that violate agreed-upon conditions?

Reconciling accessibility with control is not simply a matter of adjusting technical settings—it is a matter of redefining the ethics of data stewardship. The challenge for today and beyond is to ensure that the drive for interoperability does not eclipse the right of communities to define the terms under which their knowledge is shared, withheld, or transformed.

3. Ethics as Design or Compliance

A persistent question in data governance is whether ethics should be embedded directly into the architecture of systems—ethics by design—or treated as an external compliance requirement checked after technical specifications are finalized. The FAIR data principles were conceived primarily as a technical framework, with ethics left largely implicit, while the CARE data principles foreground ethics as a foundational element. Yet in practice, even systems that cite CARE often operationalize ethics through policy documents and consent forms, not through the technical affordances of their infrastructure.

The difference is not trivial. If ethics is an add-on, it becomes vulnerable to omission when budgets tighten, deadlines loom, or developers prioritize performance over governance. “Compliance mode” ethics tends to reduce complex cultural obligations to checklists, where the goal is to demonstrate adherence rather than to engage meaningfully with affected communities. In contrast, ethics by design would require technical choices—such as metadata schema fields, access-control mechanisms, and API endpoints—to be co-created with those communities, ensuring that governance principles are literally built into the system’s logic.

Consider a biodiversity repository that allows for both open and restricted datasets. In a compliance-oriented model, restrictions might be documented in a PDF attached to the dataset record, placing the burden on the user to read and interpret them. In an ethics-by-design model, the repository’s API would enforce those restrictions programmatically, denying certain queries unless the requester meets pre-defined criteria, and logging every access for audit. The first approach informs; the second enforces. Which is more aligned with responsible data management, and which better supports cultural data sovereignty?

Moving ethics into the design layer raises its own questions: Who decides which ethical principles are codified into the infrastructure? How are they updated as norms evolve? What happens when ethical commitments conflict with interoperability goals, funding requirements, or national data policies? If a community protocol forbids sharing specific species occurrence data, but FAIR standards encourage maximum openness, should the system be built to privilege the protocol over the standard—and if so, who takes responsibility for the trade-off?

• Integration: At what stage of system design should ethical principles be translated into technical specifications?
• Authority: Who has the legitimacy to determine which ethics are embedded and which are optional?
• Adaptation: How can ethical frameworks adapt to new technologies without losing their core intent?
• Enforceability: How do we ensure that ethical rules remain binding when data leave the original system?

Ethics by design is more than a technical challenge—it is a political stance. It demands that developers, funders, and governance bodies see ethics not as an external constraint but as an intrinsic part of system functionality. The challenge for today and beyond is to move from performative compliance toward infrastructures where respect for rights, responsibilities, and relationships is inseparable from the code itself.

C. Applied and Policy Dimensions

1. Implementing Interoperability in Biodiversity Data

Delivering interoperability in biodiversity data is rarely a purely technical matter. It plays out across semantic, technical, organizational, and legal–ethical layers, each carrying assumptions that deserve scrutiny. Treating interoperability as an engineering challenge alone often produces elegant pipelines that move data rapidly but strip away context, undermining epistemic plurality. Conversely, treating it solely as a governance issue can delay the adoption of standards that would make local data visible beyond its origin. The most sustainable implementations acknowledge that every design choice is also a governance decision.

At the semantic layer, biodiversity infrastructures rely heavily on standards such as Darwin Core and Audubon Core. These schemas allow records to be merged and queried across institutions. Yet they also embed epistemic commitments: which taxonomies are authoritative, which data fields are “mandatory,” and which forms of knowledge (seasonal indicators, cultural significance) are invisible in the model. If a concept has no field, does it effectively cease to exist for downstream users? And if so, is that an acceptable cost of interoperability?

The technical layer—persistent identifiers, APIs, and data harvesters—enables scale but amplifies mistakes. A misidentified species or overly precise coordinate can replicate across dozens of mirrors before anyone notices. Should infrastructures pause synchronization until sensitive flags are validated? Or would such delays undermine the very openness FAIR champions? The question is whether speed is an unqualified good, or whether deliberate friction is sometimes a feature, not a bug.

Organizationally, large aggregators often depend on smaller, resource-limited nodes—community archives, field stations, citizen science groups—to supply data. Yet these nodes may bear the brunt of compliance burdens without sharing equally in the benefits. When a global platform profits reputationally or financially from aggregated content, what obligations does it have to its smallest contributors? Is attribution enough, or should there be mechanisms for revenue sharing, technical support, and co-authorship in derivative analyses?

Finally, the legal–ethical layer reveals the hardest trade-offs. Systems designed for “open by default” struggle to accommodate cultural data sovereignty or conditional access rules. If a community requests withdrawal of data that meet FAIR criteria but violate local protocols, should the aggregator comply unconditionally? What if the data have already been used in peer-reviewed research? Without technical enforcement, “conditional access” becomes a polite suggestion rather than a binding rule.

Patterns of Success

• Bounded Interoperability: Core standards for global discovery, extensions for local governance and context.
• Governance-Carrying Metadata: Machine-readable consent terms and sensitivity flags that survive harvesting and mirroring.
• Tiered Access by Design: Programmatic enforcement of access levels, with logged queries for accountability.

Common Failure Modes

• Invisible Governance: Community protocols buried in PDFs that APIs ignore entirely.
• One-Way Publication: Aggregators ingest but cannot process takedown or embargo requests effectively.
• Unsafe Defaults: Sensitive geospatial data published without automatic generalization or redaction.

Critical Questions

• Who gets to decide the balance between openness and protection in global biodiversity systems?
• Can metadata schemas evolve fast enough to reflect emerging cultural, legal, and ecological sensitivities?
• Should interoperability slow down when it risks violating ethical data sharing agreements?
• How will downstream users be held accountable for breaches of responsible data management once data leave the source?

Implementing interoperability is not about choosing between FAIR and CARE—it is about resisting the temptation to let technical ease dictate ethical practice. The challenge for today and beyond is to build infrastructures where data can move across systems without forcing the values that govern them to move aside.

2. Governance Challenges

Interoperable biodiversity infrastructures are not self-governing; they are built and maintained by organizations, funded by specific sources, and guided by governance arrangements that determine what gets standardized, published, and protected. The FAIR data principles and the CARE data principles both assume that governance is essential, but they diverge on who should hold that authority and how decisions should be made. FAIR-aligned projects often rely on technical working groups and standards bodies, while CARE demands that rights-holding communities be part of the governance structure itself—not merely consulted after decisions have been made.

This divergence exposes a recurring blind spot: global infrastructures tend to centralize authority in institutions with technical capacity, while the communities most affected by data governance often have the least representation. In biodiversity networks, key governance decisions—such as what constitutes “sensitive data,” which taxonomic authorities are recognized, and how access controls are enforced—are typically made by committees far removed from the landscapes and communities the data describe. Should decision-making power scale with technical contribution, or with the degree to which one’s knowledge and ecosystems are being represented?

Representation is not just about having a seat at the table; it is about shaping the agenda and the rules of deliberation. Token participation—where community representatives are invited but given little influence over final outcomes—risks eroding trust and reinforcing extractive patterns. Genuine power-sharing means co-chairing governance bodies, setting priorities jointly, and agreeing on veto powers for decisions that contravene cultural data sovereignty or responsible data management.

Accountability is another challenge. When governance spans multiple countries and legal systems, breaches of ethical data sharing agreements can fall into a grey zone where no single authority has jurisdiction. If a dataset is misused after being downloaded from a FAIR-compliant repository, who is responsible: the original provider, the aggregator, or the end user? CARE’s emphasis on relational accountability suggests that responsibility is shared, but shared responsibility can easily become no responsibility if not clearly codified.

Key Governance Questions

• Who defines the criteria for sensitive data, and how often are these definitions reviewed?
• What proportion of governance seats are allocated to rights-holding communities versus technical or institutional actors?
• How are disputes resolved when FAIR’s openness and CARE’s protection mandates conflict?
• Should global biodiversity infrastructures be subject to independent ethical audits, and if so, by whom?
• How is downstream accountability enforced across borders when misuse occurs?

Governance challenges are not side issues—they determine the conditions under which interoperability either reinforces equity or undermines it. The challenge for today and beyond is to ensure that governance bodies are not just technically competent, but also culturally legitimate, politically accountable, and capable of adapting as knowledge systems and societal values evolve.

3. Aligning FAIR and CARE in Practice

Calls to “align” the FAIR data principles with the CARE data principles have become common in biodiversity and cultural data policy. Yet alignment is not a mechanical exercise—there is no simple mapping table where each FAIR element finds a CARE equivalent. FAIR emphasizes machine-readability, standardized formats, and broad access; CARE foregrounds cultural data sovereignty, collective benefit, and relational accountability. The question is whether these frameworks can be integrated without one diluting the other, and whether the resulting hybrid serves both technical efficiency and ethical integrity.

Some projects have attempted “CARE-aware FAIR” workflows: repositories that remain interoperable with global aggregators while embedding governance metadata, consent tags, and tiered access controls. Others have pursued “FAIR-aware CARE” models: community-controlled archives that adopt FAIR-compatible metadata for discoverability, but restrict actual data downloads to vetted users. In both cases, success hinges on sustained negotiation, not just technical fixes. Without trust and shared governance, interoperability can easily revert to business-as-usual data extraction.

There are also cautionary tales. In some biodiversity networks, CARE elements are adopted as a symbolic gesture—a few extra metadata fields or an opt-in consent statement—while default system behavior still prioritizes openness. This “CARE-light” approach satisfies funder requirements without shifting power, leaving communities with the same structural vulnerabilities. Conversely, rigidly prioritizing CARE without technical planning can isolate datasets from wider research ecosystems, reducing visibility and potential benefit for the contributing communities themselves.

The deeper challenge lies in how each framework defines “benefit” and “access.” FAIR’s benefit is typically framed in terms of efficiency, reproducibility, and innovation. CARE’s benefit is tied to equitable outcomes for specific rights-holding groups. These definitions are not always compatible. Should a dataset be considered successful if it accelerates global research but fails to deliver tangible outcomes to the community that produced it? If the answer is no, then global infrastructures must rethink their success metrics and incentive structures.

Critical Questions for Integration

• Can interoperability be redesigned so that technical integration never bypasses community governance protocols?
• What counts as “collective benefit,” and who evaluates whether it has been achieved?
• Should CARE compliance be mandatory for all datasets involving Indigenous knowledge, even if this limits their FAIR score?
• How can funding agencies reward projects that successfully maintain both FAIR discoverability and CARE protection?
• When FAIR and CARE principles conflict, what decision-making body has the legitimacy to choose between them?

Aligning FAIR and CARE in practice is not about creating a single universal standard—it is about designing processes and infrastructures that can hold space for disagreement, adapt to evolving norms, and make the trade-offs visible rather than hiding them in technical defaults. The challenge for today and beyond is to treat alignment not as a destination, but as an ongoing negotiation grounded in mutual respect and critical self-examination.

D. Future Directions

1. Designing for Epistemic Plurality

Building infrastructures that truly honor epistemic plurality requires more than appending new metadata fields or adding advisory committees. It calls for a redesign of how interoperability itself is conceptualized. Instead of treating multiple epistemologies as “edge cases” to be harmonized into a dominant standard, future biodiversity systems could be built to allow multiple classification systems, multiple authority lists, and multiple interpretations to coexist and be queried side by side. This means interoperability becomes not a flattening process, but a translation process—one that makes differences visible rather than erasing them.

Such a vision is technically possible: ontology mapping, multilingual taxonomies, and federated queries can already bridge systems without forcing premature consensus. The harder part is political: Who decides which epistemologies are included? How are conflicting interpretations displayed without privileging one over another? And how do we design interfaces so that users understand they are seeing multiple truths, not a single universal answer? These are design questions with ethical stakes, because every interface choice subtly instructs users on whose knowledge is considered primary.

In biodiversity data, for example, this could mean allowing a single species record to be annotated with both a Linnaean name and an Indigenous name, with equal prominence and linked contextual information. It could mean that ecological time-series data include both satellite-derived temperature readings and locally observed seasonal markers, without one being relegated to a footnote. But doing so raises new interoperability challenges: How do you reconcile different temporal resolutions, measurement units, or classification boundaries without distorting the source knowledge?

Designing for epistemic plurality also requires long-term commitments: to maintain multiple data models in parallel, to fund translation and interpretation work, and to train curators capable of navigating across epistemic boundaries. Without sustained investment, “plurality” risks becoming an empty label—cited in mission statements but unsupported in practice. And even when resources are available, pluralistic design can be politically contested: some institutions may resist the visibility of alternative epistemologies if they challenge established authority or commercial interests.

Key Design Questions

• What governance structures can decide which epistemologies to represent, and how to represent them?
• How can user interfaces be designed to display multiple, potentially conflicting interpretations without implying that one is “wrong”?
• What technical architectures can support parallel ontologies without excessive complexity or performance loss?
• How will long-term funding for maintaining plural infrastructures be secured, especially when they resist standardization economies of scale?

If FAIR and CARE are to coexist meaningfully, designing for epistemic plurality must be a core principle rather than an afterthought. The challenge for today and beyond is to resist the gravitational pull of single-framework dominance, and instead embrace the messiness of multiple truths—not as a problem to solve, but as a reality to sustain.

2. Policy Recommendations

Translating the integration of FAIR data principles and CARE data principles into actionable policy requires more than aspirational statements. It means setting binding requirements, funding structures, and evaluation criteria that incentivize both technical interoperability and cultural data sovereignty. Without concrete policy levers, “alignment” risks remaining a rhetorical goal while systems continue to privilege speed, scale, and openness over ethical nuance and equitable governance.

Policies should begin by recognizing that responsible data management is not only a technical matter but a rights-based one. This recognition must be embedded into the mandates of funding agencies, the terms of international data-sharing agreements, and the bylaws of standards bodies. Policies should also acknowledge that alignment between FAIR and CARE will sometimes mean accepting limits on openness in order to preserve trust, protect sensitive knowledge, or honor relational obligations.

However, prescriptive policies alone will not resolve the tensions. Enforcement and accountability mechanisms must be built in from the start. Without independent oversight, well-meaning policies can devolve into box-ticking exercises that leave harmful defaults intact. Regular ethical audits, public reporting of compliance, and clear consequences for breaches are essential to give these policies weight. Otherwise, they risk being as ephemeral as the funding cycles that produce them.

Key Policy Recommendations

• Mandatory CARE Compliance for Relevant Data: Require that any dataset involving Indigenous knowledge or culturally sensitive biodiversity information meets CARE-based governance criteria before being shared through FAIR-compliant systems.
• Ethics by Design Standards: Funders should require technical proposals to show how ethical considerations are embedded in system architecture, not just addressed in supplementary documents.
• Equitable Governance Representation: Mandate proportional representation of rights-holding communities in decision-making bodies, with authority to veto actions that violate cultural protocols.
• Benefit-Sharing Mechanisms: Establish clear channels for returning value—financial, technical, or informational—to the communities that supply the data.
• Regular Ethical Audits: Independent review of both technical and governance practices to assess adherence to FAIR–CARE alignment goals, with findings made public.
• Dynamic Consent Infrastructure: Support technical systems that allow communities to update or revoke data-sharing permissions, with changes propagated across all mirrors and derivatives.

None of these recommendations will succeed in isolation. They must be implemented together, with sustained funding and genuine co-governance. The challenge for today and beyond is to create policies that are not only aspirational, but enforceable, adaptable, and grounded in the lived realities of those whose knowledge and ecosystems are at stake.

Conclusion

The journey through the FAIR data principles and the CARE data principles is less about choosing between openness and protection, and more about confronting the assumptions that quietly shape both. FAIR has propelled unprecedented advances in data accessibility standards and interoperability, making it possible to connect datasets across disciplines, borders, and decades. CARE has re-centered the conversation on cultural data sovereignty, relational accountability, and the moral obligations of knowledge sharing. Together, they illuminate both the promise and the paradox of building a truly equitable data ecosystem.

Yet integration is neither automatic nor inevitable. The practical tensions—between speed and deliberation, universality and locality, discovery and consent—cannot be solved with a single standard or policy. They must be managed as ongoing negotiations, revisited as technologies evolve and as communities articulate new priorities. Systems that treat these tensions as inconveniences to be engineered away risk repeating the extractive patterns of the past under the banner of innovation.

Looking forward, the question is not simply how to make FAIR and CARE “compatible,” but how to design infrastructures, governance models, and funding systems that can hold multiple truths in view without collapsing them into one. This means embracing the fact that interoperability may sometimes mean translation rather than assimilation; that accessibility may sometimes mean conditionality; and that benefit must be defined by those who stand to gain or lose the most from data use.

Open Questions for Future Work

• Can global biodiversity infrastructures be designed to adapt in real time to evolving cultural and ecological sensitivities?
• What institutional incentives can ensure that ethical commitments survive beyond the life cycle of individual projects?
• How can epistemic plurality be operationalized without overwhelming technical systems or user comprehension?
• Should international data-sharing agreements include binding CARE compliance clauses, and who would enforce them?

The challenge for today and beyond is to approach data governance as a living practice rather than a fixed blueprint—one that is built with humility, maintained with vigilance, and evaluated not only by its technical performance but by the justice it delivers.