The CARE Principles for Indigenous Data Governance

The CARE Principles for Indigenous Data Governance are the four-principle normative framework — Collective Benefit, Authority to Control, Responsibility, Ethics — released by the Global Indigenous Data Alliance (GIDA) in September 2019, formally published as a peer-reviewed article in the Data Science Journal in November 2020, and intended to be applied alongside the FAIR Principles (Findable, Accessible, Interoperable, Reusable) so that data stewards #BeFAIRandCARE. The Principles propose that the political question of who the data are about and for what purpose they are collected — and the rights of the peoples to whom the data relate — be made structurally load-bearing in the design, governance, and reuse of data infrastructure, rather than left as a downstream ethics consideration after FAIR-style open-data flows have already been built. Within the make-AI-good corpus the CARE Principles are the principal Indigenous-led publication-side framework on the data-governance side of AI and data-infrastructure work — the document that the Te Hiku Media Kaitiakitanga License and the broader indigenous-AI-governance lineage anchor on, and the framework that has begun to be operationalised by mainstream scientific data infrastructures, including the Global Biodiversity Information Facility's Indigenous Data Governance Task Group and the Earth Science Information Partners network of environmental-data repositories.

Origin and drafting process

The CARE Principles were drafted at the Indigenous Data Sovereignty Principles for the Governance of Indigenous Data Workshop — a co-hosted event of International Data Week and the Research Data Alliance (RDA) Plenary on 8 November 2018 in Gaborone, Botswana — by a group of Indigenous researchers and data practitioners working under the RDA's International Indigenous Data Sovereignty Interest Group. The drafting was co-led by Stephanie Russo Carroll (University of Arizona) and Maui Hudson (University of Waikato, Aotearoa New Zealand), and the four-principle structure was iterated through a year-long open public discussion period before being formally released by GIDA in September 2019 in the form of a one-pager carrying the principles and their twelve sub-principles. The Principles were then formally published as a peer-reviewed article in the Data Science Journal on 4 November 2020 (DOI 10.5334/dsj-2020-043, CC BY 4.0) under the authorship of Stephanie Russo Carroll, Ibrahim Garba, Oscar L. Figueroa-Rodríguez, Jarita Holbrook, Raymond Lovett, Simeon Materechera, Mark Parsons, Kay Raseroka, Desi Rodriguez-Lonebear, Robyn Rowe, Rodrigo Sara, Jennifer D. Walker, Jane Anderson, and Maui Hudson. A 2024 GIDA communiqué reaffirmed and extended the stewardship roster, naming Riley Taitingfong, Jacinta Fa'alili-Fidow, Aaron Franks, Lydia Jennings, Andrew Martinez, Jacob Prehn, Cassandra Sedran-Price, Susanna Siri, Desi Small-Rodriguez, El-Shadan Tautolo, and Tahu Kukutai alongside the original co-leads as the current generation maintaining the Principles.

The four principles

The CARE framework is organised into four top-level principles, each decomposed into three sub-principles (C1-C3, A1-A3, R1-R3, E1-E3) per the canonical one-pager and the corresponding peer-reviewed exposition in the Data Science Journal.

Collective Benefit (C). Data ecosystems must be designed and operated so that Indigenous Peoples derive benefit from data — through "inclusive development and innovation", "improved governance and citizen engagement", and "equitable outcomes". The three sub-principles (C1, C2, C3) articulate what collective benefit looks like respectively in research-and-innovation ecosystems, in governance and policymaking processes, and in the distribution of material outcomes when data work produces value. The principle inverts the default open-data presumption that the benefits of data-sharing accrue to whoever can access and process the data, and reframes benefit as a flow that must be returned to the source communities.

Authority to Control (A). Indigenous Peoples' rights and interests in data — and their authority to determine governance arrangements for data about themselves, their lands, and their resources — must be recognised in the design of data infrastructures and protocols. The three sub-principles (A1, A2, A3) address respectively the recognition of those rights and interests, the use of data for Indigenous governance and self-determination, and the active role Indigenous Peoples play in governing data held by third parties. This principle is the structural anchor for the Indigenous-Data-Sovereignty register through which the broader framework is read.

Responsibility (R). Those working with Indigenous data carry an active responsibility to nurture respectful relationships with the peoples to whom the data relate — including responsibilities to expand the capability and capacity of Indigenous communities to engage with data on their own terms, and to ensure that data and the systems built on them are embedded within Indigenous languages, worldviews, and knowledge traditions. The three sub-principles (R1, R2, R3) extend the relational duty respectively to positive working relationships, to capability and capacity development, and to language and worldview grounding.

Ethics (E). Indigenous Peoples' rights and wellbeing must be the primary concern at all stages of the data lifecycle and across the data ecosystem. The three sub-principles (E1, E2, E3) address respectively the minimisation of harm and maximisation of benefit, the justice register through which historic and structural inequities in data work are read, and the responsibility-of-future-use through which present consent-to-use must be evaluated against downstream applications that are not yet visible. The Ethics principle is the one through which the framework anchors to the United Nations Declaration on the Rights of Indigenous Peoples (UNDRIP), which the operationalisation literature names as the binding international-rights register that reaffirms Indigenous Peoples' rights to control data about their peoples, lands, and resources.

Relationship to FAIR — "Be FAIR and CARE"

The CARE Principles were explicitly designed to complement the FAIR Guiding Principles for scientific data management (Findable, Accessible, Interoperable, Reusable, Wilkinson et al. 2016) rather than to oppose them. The framework's #BeFAIRandCARE hashtag and tagline names the joint stance: FAIR addresses the technical attributes of data objects themselves (whether they can be found, accessed, joined, and reused), while CARE addresses the people and purpose dimensions (who the data are about, what they are for, who controls them, what historic and structural harms are at stake in their collection and use). The 2021 Scientific Data paper Operationalizing the CARE and FAIR Principles for Indigenous data futures, co-authored by Carroll, Hudson, and three RDA / European data-infrastructure leads, documents the partnership between the RDA International Indigenous Data Sovereignty Interest Group and the RDA FAIR Data Maturity Model Working Group through which the two frameworks have been worked together in practice — naming the Nagoya Protocol, genomic databases, and the GEOME (Genomic Observatories Metadatabase) initiative as test domains for joint CARE-and-FAIR operationalisation. The pairing distinguishes CARE from earlier critiques of open-data orthodoxy that read the openness commitment itself as the problem: CARE keeps the technical FAIR machinery in play, and adds the people-and-purpose register as the prior question that determines whether and how the technical machinery applies.

GIDA and the institutional substrate

The Global Indigenous Data Alliance was formally established at a 11-12 July 2019 workshop in Oñati, Spain, convened by Maggie Walter and Desi Rodriguez-Lonebear, to bring together three Indigenous Data Sovereignty networks that had been developing in parallel through the mid-2010s: the Maiam nayri Wingara Aboriginal and Torres Strait Islander Data Sovereignty Collective in Australia; Te Mana Raraunga Māori Data Sovereignty Network in Aotearoa New Zealand; and the United States Indigenous Data Sovereignty Network. GIDA's stated mission is to "share frameworks, tools, and processes to help guide the practice of Indigenous Data Sovereignty around the globe", and the name itself — GIDA meaning guide in Basque — was chosen to reflect that guiding role. The alliance is the institutional home that has stewarded the CARE Principles since the September 2019 release, including hosting the canonical project page, coordinating with the RDA International Indigenous Data Sovereignty Interest Group, and convening the working groups behind the subsequent operationalisation and adoption literature.

Adoption and operationalisation

Since the September 2019 release the CARE Principles have moved out of the Indigenous-Data-Sovereignty literature and into mainstream scientific-data-infrastructure practice. The Global Biodiversity Information Facility (GBIF) established an Indigenous Data Governance (IDGov) Task Group to help the GBIF network — a 100-plus-country biodiversity-data infrastructure — implement the Principles through pilots aimed at making "data models, vocabularies, standards and services CARE-compliant". The 2024 Earth Science Data Repositories: Implementing the CARE Principles paper documents a parallel adoption arc in Earth and environmental data through the Earth Science Information Partners (ESIPFed) network, working in partnership with the Collaboratory for Indigenous Data Governance and GIDA. The 2024 GIDA communiqué CARE Directs Us Home — published in the context of a five-year retrospective on the Principles' uptake — surveys the field of subsequent CARE-influenced policy and infrastructure work and reasserts the Indigenous-stewardship register against incipient attempts to read the Principles as a generic data-ethics framework severable from Indigenous Peoples' rights and authority.

Position within the corpus

The CARE Principles are the corpus's principal Indigenous-led publication-side framework on the data-governance side of the make-AI-good movement, and the framework against which the corpus's existing Indigenous-AI-governance work reads. The Te Hiku Media Kaitiakitanga License — the Te Hiku Media-stewarded Māori-data-sovereignty licensing framework under which the Papa Reo te reo Māori speech-recognition corpus is held — operates in the same lineage and in fact carries forward the Te Mana Raraunga Māori Data Sovereignty Principles that are one of the three founding national networks behind GIDA, with the 2021 Taiuru-derived Māori Data Sovereignty Licences explicitly embedding the United Nations Declaration on the Rights of Indigenous Peoples in their licensing scaffolding. CARE is the international-frame normative document under which that organisation-level licensing practice sits.

As a publication type CARE fills the manifesto slot — a substantive multi-principle normative framework rather than a single-issue position statement or open letter — and sits structurally alongside the corpus's other foundational standalone-framework manifestos: the Toronto Declaration (2018), with which it shares its drafting vintage and its anchoring in binding international human-rights instruments; the Feminist Principles of the Internet, with which it shares its Global-South-rooted movement-side drafting register and its consent-and-control politics; and Oppressive A.I.: Feminist Categories to Understand its Political Effects, with which it shares the political-power register through which data colonialism and the structural-inequities-of-AI deployment are read. CARE's distinct contribution within that grouping is its Indigenous-led drafting and ongoing stewardship through GIDA, and its specific articulation of the people-and-purpose register against the data-centric register of the FAIR Principles — a contrast that gives the framework analytical purchase in scientific-data-infrastructure and AI-training-corpus contexts where rights-based manifestos drafted in legal-instrument registers have not yet found native operationalisation pathways.