CASRAI Dictionary

Category: Uncategorized

  • Data papers, software papers, and the limits of CRediT

    The 14 roles of CRediT were designed against the model of a conventional research article reporting empirical work: a study with a hypothesis, a method, data, analysis, and a written argument. Data papers and software papers fit this model awkwardly. A data paper describes a dataset; a software paper describes a piece of software. The intellectual contribution is the artefact itself, not the prose around it. The CRediT roles, applied to these papers, produce statements that are technically valid but substantively misleading. This post catalogues the friction and suggests where the taxonomy could be extended.

    What a data paper actually is

    A data paper, as the genre has developed in venues like Scientific Data, Earth System Science Data, GigaScience, and the data-paper streams of disciplinary journals, is a peer-reviewed description of a dataset: its provenance, its collection method, its quality, its access conditions, and its potential reuse. The dataset itself lives in a repository with its own DOI; the data paper provides the citable, peer-reviewed scholarly record that the dataset exists, that it was collected with rigour, and that it is fit for reuse.

    The intellectual labour behind a data paper is mostly not in the paper. It is in the years of fieldwork or instrument operation that produced the data, the protocols that ensured comparability across collection events, the curation work that turned raw observations into a structured deposit, the documentation that lets a stranger understand what the data mean. The paper is a summary record of that work.

    Where CRediT falls short for data papers

    Three friction points. First, Investigation and Data curation bear most of the load and they are not differentiated finely enough. A field ecologist who spent years collecting samples, a lab technician who processed them, a data manager who normalised the schema, and a metadata specialist who wrote the documentation are all plausibly Investigation or Data curation; the roles do not distinguish them. The result is that two papers with very different actual contributorship patterns can have identical-looking CRediT statements.

    Second, Resources overlaps with Investigation in a confusing way. A data paper describing a long-term ecological observatory has a Resources contribution (the observatory itself) that is distinct from the per-sample Investigation. CRediT does not currently cleanly separate “provided the infrastructure that produced the data” from “provided the samples that went into the data.”

    Third, Writing – original draft is often the smallest contribution, not the largest, and assigning it Lead can misrepresent the contribution structure. The person who wrote the paper is often a relatively junior team member, not the senior person whose intellectual contribution was the protocol and the multi-year campaign.

    Software papers and the JOSS model

    Software papers, exemplified by the Journal of Open Source Software (JOSS), face an analogous problem from a different direction. A JOSS paper is short — often under 1,000 words — and is paired with a peer-reviewed software repository. The intellectual contribution is the software: its design, its implementation, its tests, its documentation, its maintenance over time. The paper is a stub.

    JOSS itself uses CRediT for its papers and has done so since 2020. The community has converged on a set of mappings:

    • Conceptualization covers software design and architectural decisions.
    • Software covers implementation. This is the central role for most JOSS contributors.
    • Validation covers testing, both unit tests and validation against reference implementations.
    • Methodology covers the algorithmic content, where the software implements a non-trivial method.
    • Writing – original draft covers the paper itself. The README, the developer documentation, and the user docs are also writing work, but they are not the JOSS paper.
    • Supervision covers project leadership; Project administration covers maintenance and coordination.

    The friction in this mapping is that the Software role is overloaded. It conflates the initial implementation, ongoing maintenance, bug-fixing, refactoring, and tooling. A contributor who implemented the core algorithm and a contributor who maintains the CI/CD pipeline both get “Software” with no further distinction. For long-lived software with many contributors over years, the role assignment ends up giving everyone Software (lead/equal/supporting) and the differentiation lives in the GitHub commit history, not in CRediT.

    The FAIR4RS angle

    The FAIR4RS Principles for research software, finalised in 2022, set out what FAIR means for software: findable, accessible, interoperable, reusable. They explicitly acknowledge that software citation needs a richer model than data citation, because software has versions, dependencies, and ongoing development that data typically does not.

    FAIR4RS implies, though does not directly require, a richer contributorship taxonomy for software. The Software Citation Implementation Working Group has been chewing on this for several years. Their working position is that CRediT remains the right vocabulary for software paper contributorship, but that the software repository itself should carry its own contributor metadata using a complementary scheme — typically CITATION.cff with extended fields — that captures the per-version, per-component contributorship that CRediT cannot.

    The mapping problem

    For data papers and software papers, the operational reality is that two parallel records exist: the paper’s CRediT statement and the dataset or software repository’s contributor metadata. They overlap but do not align cleanly. The dataset DOI and software DOI live in DataCite; the paper DOI lives in Crossref; the relations between them are declared in the metadata but not always reciprocally.

    The CASRAI research outputs domain tracks the mapping conventions in current use. Our recommendation, for now, is that data papers and software papers should publish a CRediT statement covering the paper’s contributorship and should additionally publish a richer contributor metadata file with the dataset or software, using CRediT roles plus the disciplinary-specific extensions that have emerged.

    Possible extensions

    Three extensions would meaningfully improve the situation. First, sub-roles within Software: an extended taxonomy with implementation, testing, documentation, maintenance, and integration as sub-roles would give a software paper a more truthful contributorship statement. This work has been drafted by the FORCE11 software citation working group but not formally proposed as a CRediT extension.

    Second, distinguished Investigation roles for data papers: collection, processing, curation, documentation as sub-roles of Investigation and Data curation would let a data paper describe its contributorship more faithfully. The challenge here is keeping the taxonomy usable; an over-elaborate vocabulary loses adoption.

    Third, artefact-level role assignments: the current CRediT statement applies at the paper level. For a paper that describes a dataset and a software package, it might be more useful to have role assignments at the artefact level (paper, dataset, software each get their own statement) with cross-references. This would require schema work in Crossref, DataCite, and ORCID.

    What to do now

    For authors of data papers, the practical advice is: use CRediT for the paper; deposit a complementary contributors.json with the dataset that captures finer-grained roles; cross-reference the two in the related-identifier blocks. For authors of software papers, use CRediT for the paper and CITATION.cff for the repository, with the CFF carrying the rich per-component contributor data. The CASRAI data and software papers guide has worked examples.

    For the CRediT stewardship group, the recommendation is to prioritise the data-paper and software-paper mapping problem in the v2026.3 revision discussion. The friction is real, the workarounds are working but ugly, and the taxonomy will be strengthened by a thoughtful extension.

    Related dictionary entries

  • Paper mills and tortured phrases: the integrity crisis in 2026

    The scholarly-publishing integrity ecosystem ended 2025 with the highest retraction rate ever recorded and the clearest evidence yet that industrial-scale fraud is structurally embedded in the literature. The numbers are sobering: Retraction Watch’s database crossed 60,000 entries in 2025; Hindawi/Wiley alone retracted over 11,000 papers across 2023-2024 following paper-mill detection; the Problematic Paper Screener now flags new manuscripts at a rate that strains journals’ capacity to investigate. This post maps the current threats, the detection tooling that has matured, and the United2Act coordination work that is beginning to produce a coherent industry response.

    Paper mills: the supply side

    A paper mill is a commercial operation that fabricates manuscripts and sells authorship slots on them. The mills emerged at significant scale around 2010-2012, driven by promotion-and-tenure incentives in jurisdictions where publication count is a hard quantitative requirement (early-career clinical researchers in some countries face explicit per-promotion-step publication quotas). The mills industrialised what individual fabrication had done for decades.

    The 2022-2024 Hindawi crisis (Wiley’s acquired open-access portfolio was infiltrated at scale, leading to 11,000+ retractions and the closure of several journals) made the systemic nature visible. The Hindawi pattern was: mill-generated manuscripts submitted to special-issue calls in low-rigour journals, peer-reviewed by mill-affiliated reviewers in coordinated networks, published, and used for career advancement. The breakdown was multifactorial: high-volume special-issue calls without sufficient editorial oversight; reviewer networks that the journal could not detect were coordinated; a financial incentive structure that rewarded throughput.

    The 2024-2025 response was substantial. Wiley shut down the Hindawi brand, retracted at scale, and rebuilt its peer-review controls. Other publishers running similar special-issue programmes audited and tightened. The COPE-led United2Act initiative (United2Act for paper mills, launched 2023) produced industry-wide commitments to detection cooperation, transparent retraction practices, and improved reviewer verification.

    Tortured phrases: a detection lever

    The tortured phrases concept, coined by Guillaume Cabanac and Cyril Labbé in 2021, was a methodological breakthrough. A tortured phrase is a clumsy paraphrasing of a standard technical term, typically introduced by attempting to evade plagiarism detection by automatic word substitution. “Counterfeit consciousness” for “artificial intelligence,” “haphazard backwoods” for “random forests,” “fake neural organization” for “artificial neural network.” Once recognised, tortured phrases are a reliable signal of mill involvement, because no human author working in their field would write “haphazard backwoods” when they meant random forests.

    Cabanac and Labbé’s Problematic Paper Screener (PPS) operationalises tortured-phrase detection at scale. The PPS continuously scans the published literature against a curated dictionary of tortured phrases, flagging papers that contain them. By 2026 the PPS has flagged over 14,000 papers; many have been retracted, more are under investigation, and a substantial subset will likely remain in the literature without action because the journals are unresponsive or defunct.

    The PPS is open infrastructure (the dictionary is public, the methodology is published, the flagged papers are listed). It has been criticised for false positives (some flagged papers turn out to have innocent explanations, e.g., automated translation from a non-English original) but the precision is high enough that an editor receiving a PPS flag should treat it as a serious signal warranting investigation.

    Image manipulation

    The other major detection front is image manipulation, particularly in life-science papers where Western blots, microscopy images, and gel electrophoresis are routinely fabricated by duplication, splicing, or AI generation. Elisabeth Bik’s catalogue of image-duplication cases has been the canonical reference for over a decade. The 2022-2024 development was the deployment of automated image-similarity tools (Imagetwin, Proofig) by major publishers; by 2025 most large publishers run automated image screening on every submission.

    The 2025 escalation is AI-generated images. A diffusion-model-generated Western blot is more difficult to detect than a duplicated one because there is no source to find. The detection community has begun work on AI-generated-image detection but the arms race is genuinely real, with no settled tool. The current best practice is to require raw data deposition (the original blot scan, the unprocessed microscopy stack) alongside the published image, with image-manipulation tools running on both. Several Cell Press and EMBO journals now require this for all life-science submissions.

    Citation cartels

    Citation cartels are coordinated networks of authors who systematically cite each other to inflate their citation counts and journal impact factors. The classic cartel pattern is journal-level: a journal’s editorial board reciprocally cites other journals’ editorial boards, all benefiting from the inflated cross-citation. The author-level pattern is similar: a network of researchers in adjacent specialties cites each other across many papers.

    Detection is statistical: cartels show citation patterns that are sharply non-random in the citation graph. The 2023-2024 work by Albers Mohrman and others operationalised the detection at the journal-citation-network level; Clarivate has begun excluding cartel-implicated journals from the JCR. The author-level cartels are harder to act against, but the existence of the signal is becoming part of the institutional-integrity toolkit.

    The retraction infrastructure

    Retraction has historically been slow, opaque, and inconsistently practiced. The 2022 NISO recommended practice on retraction (NISO RP-45-2022) and the 2024 Crossref retraction-metadata revisions have begun to change this. A retracted paper now carries structured machine-readable metadata about the retraction reason, the implicated parties, and the relationship to other papers; downstream services (PubMed, Google Scholar, Scopus, citation databases) consume the metadata and surface retraction notices alongside the paper.

    The remaining gap is the unretracted-but-suspect paper. A paper flagged by the Problematic Paper Screener but never investigated by the journal sits in the literature unmarked. The 2024 COPE-led discussion of expressions of concern as an interim status (the paper is under investigation but not yet retracted) is one direction. A more radical proposal, now being piloted by several preprint servers and one or two journals, is to surface the PPS flag directly on the article landing page even before the journal acts, with a clear distinction between “flagged by automated screener” and “retracted by publisher.”

    The United2Act response

    United2Act, launched in 2023 with COPE and STM coordinating, brought publishers, researchers, integrity offices, and regulators together to address paper mills. The 2024 United2Act communique committed signatories to: cooperate on detection (sharing reviewer-misconduct signals across publishers); standardise retraction practices; improve reviewer verification; coordinate with institutions on consequences for authors of fabricated papers.

    The 2025 work has been operational: the COPE/STM joint paper-mill database (publishers can submit suspect manuscript signatures and the database flags coincidences); reviewer-verification protocols (ORCID iD plus institutional email plus referee history); coordination with national integrity offices in jurisdictions where paper-mill commissioning is concentrated.

    The honest assessment is that United2Act has bought the industry better coordination but has not solved the structural incentive problem. As long as researchers face quantitative publication requirements for promotion, the demand for fabricated authorship slots will exist. The longer-term fix is on the responsible-assessment side (see our responsible-assessment domain); the integrity-side work is harm reduction.

    COPE flowcharts: the per-case operational layer

    The COPE flowcharts, maintained and updated by the Committee on Publication Ethics, are the operational toolkit for editors handling suspected misconduct. The flowcharts cover (among many) plagiarism in a submitted manuscript, plagiarism in a published article, redundant publication, fabricated data, undisclosed conflict of interest, undisclosed AI use, image manipulation, authorship disputes, paper-mill suspicion, and citation manipulation.

    An editor confronted with a suspect submission in 2026 should pull the relevant COPE flowchart, follow the documented procedure, and document the decision trail. The flowcharts are not a substitute for editorial judgement, but they are an audit-defensible baseline. The 2024-2025 COPE updates added flowcharts specifically for AI-assisted fabrication, paper-mill suspicion based on tortured-phrase detection, and image-manipulation findings from automated tools.

    What to do at the institutional level

    For an institutional research-integrity office in 2026, the practical priorities are: (1) monitor your own institution’s authors against the PPS and the Retraction Watch database; (2) integrate retraction-metadata feeds into your CRIS so you can detect when your authors’ papers are retracted elsewhere; (3) participate in United2Act or its national-level analogues; (4) commit publicly to following COPE flowcharts and document decisions; (5) work with your promotion-and-tenure committees to remove the pure-count incentives that fuel the demand side. The research-integrity domain at CASRAI maintains the institutional-integrity playbooks.

    Related dictionary entries

    References

    Cabanac, Labbé, Magazinov, Tortured phrases: A dubious writing style emerging in science (2021 preprint and follow-up papers). Bik et al., The Prevalence of Inappropriate Image Duplication in Biomedical Research Publications (mBio, 2016). Else and Van Noorden, The fight against fake-paper factories that churn out sham science (Nature, 2021). COPE, Paper mills – research, action plans, and resources (2023, updated 2024). United2Act, Joint Communique on Paper Mills (2023).

  • Sustainable laboratory operations: LEAF, My Green Lab, and the carbon footprint of research

    The carbon footprint of research is unusually large per worker. A typical wet lab consumes 3-10 times the energy of an equivalent office space; a single ultra-low-temperature freezer running 24/7 uses as much electricity as an average household; single-use plastics in life-science labs alone are estimated globally at 5.5 million tonnes per year. A research-intensive university’s Scope 1 and 2 emissions sit primarily in lab buildings; its Scope 3 sits primarily in travel and procurement. The sustainability conversation in 2026 has moved from awareness to operational programmes, with two frameworks dominating: LEAF (the Laboratory Efficiency Assessment Framework developed at UCL) and My Green Lab certification. This post is a practical tour.

    Why labs are different

    Office sustainability programmes (LED lighting, paper recycling, energy-efficient computing) translate poorly to laboratories because the energy intensity is in equipment that cannot simply be switched off. A freezer holding biological samples cannot be turned off at night; a fume hood cannot be reduced to standby; a sequencer cannot run at half power. Lab sustainability is therefore primarily a question of which equipment runs and how it is operated, not whether it is on.

    The corollary is that lab sustainability programmes need a different vocabulary and a different evidence base than office sustainability. Both LEAF and My Green Lab were designed in response to this; both have been validated empirically over the 2018-2024 period.

    LEAF

    The Laboratory Efficiency Assessment Framework, developed at University College London by Martin Farley and colleagues, is a self-assessment-and-certification framework structured at three levels (Bronze, Silver, Gold). LEAF’s strength is that it is operationally specific: each level lists discrete actions a lab can take, and the actions are tied to estimated energy and waste impacts based on UK lab benchmarks.

    The LEAF Bronze level covers basics: freezer temperature optimisation, sash management on fume hoods, equipment shutdown protocols, recycling, lighting, water conservation. Silver adds: research-life-cycle assessment, supplier engagement, training, advocacy. Gold adds: integrated sustainability planning, leadership in the institution’s sustainability programme, mentorship of other labs.

    By 2026 LEAF is in use at over 1,000 institutions globally, with strong concentration in the UK (where it was developed and has the strongest institutional backing). The LEAF self-assessment is free; the certification process involves institutional review. The framework is open and has been adapted to several national contexts.

    The LEAF impact data

    The 2023 UCL study of LEAF Bronze-certified labs found average energy reductions of 5-15% versus baseline, primarily from freezer optimisation and fume-hood sash management. The 2024 follow-up at LEAF Silver labs found additional 5-10% reductions and significant reductions in single-use plastic consumption through supplier engagement. The data underwrite the framework: this is not aspirational, it is documented.

    My Green Lab

    My Green Lab is a US-based non-profit operating a complementary certification programme that has dominated the North American market and increasingly the international one. My Green Lab Certification (MGLC) is built around a survey assessment in 14 categories (energy, water, waste, green chemistry, purchasing, training, etc.) with a numerical score and an annual recertification cycle.

    My Green Lab also operates the ACT Label (Accountability, Consistency, Transparency) for lab products: a vendor-supplied environmental-impact rating for an individual product (pipette tip, plate, reagent) covering its energy, water, packaging, and chemical inputs. The ACT Label is in widespread use across major lab suppliers (Eppendorf, Thermo Fisher, Bio-Rad, NEB) and has become a discriminator at the procurement stage for sustainability-minded labs.

    By 2026 My Green Lab claims over 4,000 certified labs across more than 70 countries; the certified-lab cohort is concentrated in pharmaceutical and biotechnology industrial research, with growing academic uptake.

    Choosing between LEAF and My Green Lab

    Most institutions can choose one and stick with it. LEAF is more prescriptive (the levels list specific actions); My Green Lab is more diagnostic (the survey identifies areas for improvement and tracks scores over time). LEAF has stronger UK and European institutional backing; My Green Lab has stronger US and pharmaceutical-industry backing. An institution coordinating with a major US pharmaceutical partner is likely better off with My Green Lab; one coordinating with UKRI or EU funders is likely better off with LEAF.

    Some institutions run both. The duplicate-overhead cost is moderate (the underlying lab practices are largely the same; the documentation is different) and the dual recognition can be useful.

    Freezer management: the biggest single lever

    The single most impactful intervention in most labs is freezer management. A typical -80°C ultra-low-temperature freezer consumes 16-22 kWh/day. Switching to -70°C set point (acceptable for most stored samples per the 2018 Gail Phébré et al. validation) cuts consumption by 30-40%. Combining with a sample-inventory audit (most labs have 10-30% of freezer contents that are unused, unlabelled, or duplicated and can be discarded) frees space and avoids new freezer purchases.

    The UCL Race to Zero trial in 2022-2023 had 53 labs switch -80°C freezers to -70°C; the energy savings were as predicted and no sample integrity issues were reported across 12+ months of follow-up. This is now standard guidance in both LEAF and My Green Lab.

    Single-use plastics

    The wet-lab single-use-plastic flow is enormous and largely necessary (sterility, contamination control, reproducibility). The mitigation in 2026 has two prongs. First, vendor switching: products with ACT Labels in the better tiers (post-consumer recycled content packaging, take-back programmes, reduced primary packaging) materially reduce flows. Second, recycling streams: rigid PE/PP lab plastics (tip boxes, conical tube racks) are recyclable in dedicated lab-plastic streams operated by several vendors. The recycling capture rate has grown substantially since 2022.

    The remaining hard problem is contaminated plastics (anything that has touched biological or chemical materials and cannot be cleanly recycled). The mitigation is procurement-stage: smaller tip volumes, single-use serological pipette redesigns with less plastic per unit, reusable glassware where appropriate.

    Sustainable HPC

    High-performance computing is the fastest-growing emissions source in many research universities. A modern HPC cluster running 24/7 at high utilisation has a substantial Scope 2 footprint; AI-training workloads in particular have caused HPC electricity consumption to grow sharply since 2022.

    The mitigations in 2026 include: power-aware job scheduling (running flexible jobs when grid carbon intensity is low, e.g., when wind generation is high); efficiency-first allocation (prioritising jobs that have demonstrated CPU/GPU efficiency); ML-model-efficiency policies (preferring smaller, more efficient models where they suffice); reporting emissions per project in the same way that we report compute hours.

    The Green Algorithms tool and the CodeCarbon Python library let researchers estimate emissions per analysis. UKRI and the EU’s HORIZON programme now ask researchers to report estimated emissions in proposals for compute-heavy projects.

    Conference travel: the Scope 3 elephant

    Academic conference travel is, for most research-intensive universities, the single largest Scope 3 emissions category. A round-trip transatlantic flight emits roughly 1-2 tonnes of CO2 per passenger; an academic with a typical conference cadence can easily account for 5-10 tonnes/year of travel emissions, which dwarfs everything else they personally consume.

    The 2020-2024 pandemic enforced a partial shift to virtual conferencing; the post-pandemic settlement has not held. By 2026 conference travel is largely back to pre-pandemic levels, though with somewhat more hybrid options. The frameworks that have emerged include: institutional travel-budget caps with carbon-equivalent accounting; conference-clustering (attending one trip with multiple events rather than several separate trips); flight-free regional conferences (the UK Reproducibility Network’s flight-free Easter conference, the European Geosciences Union’s hybrid format); and proportional-attendance models in which junior researchers attend in person while seniors attend virtually.

    The conference travel emissions conversation is genuinely difficult because there are real career and equity costs to reducing in-person attendance. The current best practice is to count, declare, and make trade-offs visible, rather than to impose a top-down quota.

    Scope 1, 2, 3 in research-org context

    For a research institution: Scope 1 is direct (campus heating fuel, owned vehicles); Scope 2 is purchased energy (electricity, district heating); Scope 3 is everything else (travel, procurement, commuting, waste, investments). For a typical research-intensive university, Scope 3 is 70-90% of total emissions, with travel and procurement dominating. The implication is that a serious sustainability programme must address Scope 3 procurement (sustainable lab purchasing) and Scope 3 travel (conferences and fieldwork), not just on-campus operations.

    The sustainable-research domain at CASRAI tracks framework adoption and institutional case studies; the research carbon footprint entry walks through the standard accounting methodology adapted for research organisations.

    Related dictionary entries

    References

    Farley et al., LEAF: a tool for laboratory sustainability assessment (UCL technical report, 2019, updated 2023). My Green Lab, 2024 Certification Standard (current version). Urbina et al., Labs should cut plastic waste too (Nature, 2015, the foundational plastics paper). Lannelongue et al., Green Algorithms: Quantifying the carbon footprint of computation (Advanced Science, 2021). UCL Race to Zero, Freezer temperature transition report (2023).

  • Why the next CRediT version should include ‘AI assistance’ as a role

    The 14 roles of CRediT were designed in 2013-2014 with a model of contribution that did not include large language models or generative AI systems. A decade on, the taxonomy is robust and widely adopted, but the AI question is hard to ignore. This post makes the case — tentatively, and with attention to the counter-arguments — that the next CRediT revision should add a 15th role explicitly covering AI assistance. We are publishing it here to invite community pushback before any formal proposal goes to the CRediT stewardship group.

    Why this question is not solved by disclosure alone

    The current consensus around generative AI in scholarly authorship rests on two pillars: AI cannot be a co-author (the ICMJE 2023 position), and AI use must be disclosed in a structured declaration. CASRAI agrees with both. They do not, however, resolve the question of how AI assistance shows up in CRediT.

    A worked example. Suppose a paper has four authors. Author A wrote the first draft with substantial assistance from a large language model, which she prompted, edited, fact-checked, and revised. Author B ran the formal analysis using an AI-assisted statistical-discovery tool that proposed model specifications. Author C generated several of the figures using a GenAI visualisation tool. Author D supervised. Each used AI; each used it differently; each took human responsibility for the output. How does the CRediT statement represent this?

    Under current CRediT, AI use is invisible. Author A gets Writing – original draft (lead). Author B gets Formal analysis (lead). Author C gets Visualization (lead). Author D gets Supervision. The AI assistance shows up only in the publisher-mandated AI disclosure, which is a free-text field in the methods or acknowledgements. The structured contributorship record has no place for the granular fact that AI was a tool in each of those role-discharges.

    The proposed 15th role

    The draft scope we are testing is this:

    AI assistance. The use of artificial-intelligence systems, including generative AI, machine-learning models, and automated analytical tools, in the production of the work. Includes prompt engineering, model selection, validation of AI output, and human verification of AI-generated content. Does not include use of AI as a routine tool (e.g., grammar checkers, citation-formatting tools) below a disclosure threshold defined by the publisher.

    The role would carry the standard degree-of-contribution qualifier. A human author whose primary contribution was prompting and verifying an AI system would be marked Lead for AI assistance; a co-author who occasionally checked AI outputs would be Supporting. The role would not be a substitute for the existing roles — the human who used AI for the first draft still gets Writing – original draft — but it would add the structured fact that AI was involved.

    The arguments for

    First, structured disclosure is more useful than prose disclosure. A free-text AI declaration cannot be queried, cross-referenced, or aggregated. A CRediT-style structured role can. Integrity offices investigating a fabrication can query for papers with AI assistance roles; funders tracking AI use in grant outputs can roll up the data; bibliometric studies can analyse patterns. None of this is possible with the current free-text disclosure.

    Second, granularity matters for accountability. Knowing that a paper used AI is less useful than knowing which contributor used AI for which task. The CRediT role assignment makes the accountability specific. If a fabricated reference appears in the introduction, the question of who is responsible for verifying it has a structured answer.

    Third, the boundary is becoming a fiction. Modern statistical workflows include AI components (autoML, AI-assisted exploratory analysis); modern writing workflows include AI components (Copilot for prose, Claude for editing); modern visualisation workflows include AI components. The pretence that these are separable from the role they support is increasingly hard to maintain. If AI is being used to discharge a role, the role assignment should say so.

    The arguments against

    Three serious counter-arguments deserve engagement.

    First, the scope-creep concern. CRediT has held to 14 roles deliberately. Each addition raises the cognitive load on authors filling out the statement, increases the integration burden on publishers, and risks the taxonomy becoming unusable through over-specification. The argument from Liz Allen and the original CRediT designers has been that the taxonomy gains its value from being small enough to use.

    Second, the boundary problem. What counts as AI assistance? A grammar checker is plausibly AI; a citation formatter increasingly is; a search engine ranking results by relevance certainly is. If every modern research tool counts as AI, the role becomes meaningless. A workable scope requires a non-trivial threshold (the draft language above gestures at “below a disclosure threshold defined by the publisher”), and that threshold is hard to define without ending up with either everything or nothing.

    Third, the disclosure-versus-contribution distinction. CRediT is a contributorship taxonomy. AI is not a contributor — that is the settled position. Adding an AI role to CRediT risks blurring this. The alternative is to keep AI in a separate disclosure form, structurally similar to a competing-interests declaration or a funding statement, rather than in the contributorship statement.

    A possible middle path

    The middle path is to keep CRediT at 14 roles and to define a parallel AI assistance declaration with comparable structure: a controlled vocabulary of AI-use types, a per-contributor breakdown linked to ORCID iDs, a model-and-version field, and a verification statement. This would sit alongside CRediT in publisher submission systems and JATS XML, rather than inside it.

    This is closer to where the current publisher disclosure forms are heading, and it preserves the conceptual clarity that CRediT roles describe what humans did, while a separate declaration describes what AI tools were used. We are increasingly inclined to recommend this path, with the caveat that the disclosure must be structured to the same standard as CRediT — not free-text, with controlled vocabularies, deposited to Crossref, and surfaced on ORCID.

    What the CRediT stewardship group should do next

    Three concrete steps. First, run a structured community consultation through 2026 on whether to add AI assistance as a 15th CRediT role, with the alternative being a parallel structured declaration. The CRediT governance page outlines the consultation process. Second, in parallel, draft the data model for a parallel AI assistance declaration so that the comparison is concrete and not abstract. Third, coordinate with NISO on whether either option requires a revision to Z39.104.

    The decision is not urgent in the sense that the integrity system is failing today; the existing disclosure forms work, badly. It is urgent in the sense that every year of delay produces another year of unstructured AI-use data that cannot be aggregated or analysed, which makes the eventual transition harder.

    Related dictionary entries

  • NSPM-33 disclosure: what US researchers must report in 2026

    National Security Presidential Memorandum 33 (NSPM-33), signed in January 2021, directed US federal research funding agencies to strengthen and harmonise disclosure requirements for federally funded researchers. Five years later the implementation has stabilised across NIH, NSF, DOE, DOD, NASA, USDA, and the other major science agencies, with the CHIPS and Science Act of 2022 having added enforcement teeth and the 2024 Research Security Programs Standard Requirement having added institutional-level obligations. This post is the practical 2026 compliance map for US-funded researchers.

    The shape of NSPM-33 in 2026

    NSPM-33’s core mandate is straightforward: a federally-funded researcher must disclose all support they receive (financial, in-kind, or in the form of positions, appointments, or affiliations) so that the funding agency can identify potential conflicts of commitment, undisclosed foreign components, or scientific overlap. The disclosure is made at proposal stage and updated throughout the project’s life.

    The five years of implementation have produced two important refinements. First, the common disclosure forms: NIH’s Other Support format, NSF’s Current and Pending (Other) Support, and parallel formats at other agencies have been substantially harmonised under the NSPM-33 implementation guidance. By 2026 a researcher can largely produce one structured disclosure record (typically in SciENcv format) and have it serve all federal agencies. Second, the structured-data submission: the agencies now require disclosure forms in machine-readable format with ORCID linkage, not as free-form PDFs.

    What must be disclosed

    The 2026 disclosure scope at the major agencies covers, at a minimum:

    • All ongoing and pending research support (federal, non-federal, and foreign).
    • All in-kind support of significance (laboratory space, equipment access, personnel time).
    • All positions and appointments (professorships, visiting positions, advisory roles, board memberships) regardless of whether they are paid.
    • All consulting arrangements above a defined threshold (typically a few thousand dollars per year, but agency-specific).
    • Foreign government talent recruitment programme participation (see below).
    • Patents and patent applications related to the funded research.
    • Sponsored or paid travel above defined thresholds.
    • For NIH specifically, all support for research effort regardless of how titled.

    The Current and Pending Support form (NSF terminology) and the Other Support form (NIH terminology) are the canonical artefacts. They are populated by the researcher at the proposal stage and re-verified at the just-in-time (JIT) request stage if the proposal is funded.

    The foreign-component question

    The single most consequential 2021-2024 enforcement focus was undisclosed foreign components. A foreign component is any significant scientific element of a project performed outside the United States by any source of funding, including foreign collaborator efforts even if not separately funded.

    NIH’s foreign-component disclosure rule existed before NSPM-33 but was inconsistently enforced. Post-NSPM-33 the enforcement has been substantial: dozens of researchers had grants terminated or returned, and several criminal cases proceeded for fabricated disclosures. The 2023-2024 cohort of cases clarified the threshold: an undisclosed foreign-funded position, a foreign-government talent-recruitment-programme membership, or a substantial unreported collaboration with a foreign laboratory are all material non-disclosures with grant-termination and criminal consequences.

    In 2026 the practical rule is conservative: if you have any affiliation, position, support, or significant collaboration outside the US that overlaps in time with your federal-funded project, disclose it. The cost of over-disclosure is filling in more forms; the cost of under-disclosure has become very high.

    Foreign Talent Recruitment Programmes

    The Foreign Talent Recruitment Programme (FTRP) category was sharpened by Section 10632 of the CHIPS and Science Act of 2022, which required agencies to prohibit federally-funded researchers from participating in malign FTRPs. The 2024 implementation guidance defined a malign FTRP as one that involves transfer of intellectual property, transfer of laboratory resources, or compensation contingent on outcomes that benefit a foreign government’s national interests, among several other criteria.

    The category is narrower than the original 2018-2021 “China Initiative” framing might have suggested. Participation in a non-malign FTRP (a competitive postdoctoral programme, an academic exchange visit, an honorary professorship) is not prohibited but must be disclosed. Participation in a malign FTRP is prohibited for federally-funded researchers and must be terminated as a condition of receiving federal funding.

    The institutional-side burden under the 2024 Research Security Programs Standard Requirement is substantial: institutions over a defined funding threshold must implement a research security programme with training, conflict-of-interest screening, foreign-collaboration approval, and ongoing monitoring. The standard requirement specifies the elements; institutions implement them with their own policies.

    The reporting workflow in practice

    The 2026 workflow for a federally-funded researcher at a US institution typically looks like:

    1. SciENcv profile. Maintain a current SciENcv profile with all positions, appointments, and support. SciENcv (Science Experts Network Curriculum Vitae) is the federal-government-supported tool and produces the structured-data formats accepted by NIH, NSF, and other agencies.
    2. Proposal-stage disclosure. Export the relevant disclosure form from SciENcv at proposal preparation. Verify with the institution’s sponsored-research office before submission.
    3. JIT update. For NIH, re-verify Other Support at JIT request. Any changes since proposal submission must be reported.
    4. Award updates. Any new support, position, or appointment acquired during the award must be reported to the agency. NIH’s threshold is “significant changes”; in practice, disclose anything that would have been on the original form.
    5. Annual progress reports. RPPR and other annual reporting captures updated Other Support and current-and-pending. Treat this as a real update, not a copy-paste.
    6. Final reports and closeout. Disclosure obligations continue through closeout.

    Institutional research security programmes

    The 2024 Research Security Programs Standard Requirement obligates institutions over the $50M annual federal-research threshold to operate a research security programme covering: cybersecurity training, foreign-collaboration approval workflow, conflict-of-interest and conflict-of-commitment training, export-control compliance, and ongoing monitoring of researchers’ disclosures against external data sources.

    The institutional layer matters because most disclosure failures are not fraud; they are inadvertent omission by researchers who did not realise an affiliation was disclosable. A well-functioning research security programme acts as a backstop, with regular reminders, training, and a pre-submission review that catches omissions before they become non-disclosures of consequence.

    The CASRAI funder-mandate guide covers the agency-specific disclosure requirements with current links; the research-security domain tracks the cross-agency policy harmonisation.

    What’s still uncertain

    Three areas remain in active interpretation in 2026. First, the treatment of dual-affiliated researchers: a researcher with a tenured position at a US institution and a part-time appointment at a non-US institution must disclose both, but the threshold for the non-US appointment counting as a foreign component is fuzzy in practice. Second, the scope of the conflict-of-commitment definition: an unpaid advisory role at a foreign institution may not count as support but does count as commitment; the agencies vary in how they treat this. Third, the retroactive application: disclosure failures discovered years after the funded work was completed have been treated with substantial inconsistency, with some cases pursued criminally and others handled administratively.

    For researchers, the safe path is conservative disclosure, current SciENcv maintenance, and proactive consultation with the institution’s sponsored-research office whenever an affiliation or support is ambiguous. The compliance cost of asking is low; the cost of under-disclosure that surfaces later is potentially career-ending.

    Related dictionary entries

    References

    NSTC Joint Committee on the Research Environment, Guidance for Implementing National Security Presidential Memorandum 33 (January 2022). NIH Office of Extramural Research, Notice of Information: Updates to Other Support (NOT-OD-22-150 and subsequent updates). CHIPS and Science Act of 2022, Section 10632 (Foreign Talent Recruitment Programs). OSTP, Research Security Programs Standard Requirement (July 2024). NSF, Proposal and Award Policies and Procedures Guide (current version).

  • Crossref’s grant-linking initiative and CRediT: a 2025 status

    Crossref’s Grant Linking System (GLS), in development since 2019 and in steady production since 2022, has quietly become one of the most useful bits of plumbing in scholarly metadata. Its 2025 expansion — covering more funders, more grant metadata, and tighter integration with Crossref’s content-registration deposit schema — is worth a closer look, particularly for anyone integrating CRediT contributorship with funding attribution. This post walks through what GLS does, what changed in 2025, and how a CRediT integrator should consume the data.

    What GLS is, briefly

    A grant, as a thing in the world, has a funder (an organisation that paid), one or more awardees (people and institutions), a title, a project description, an amount, a duration, and a set of outputs (papers, datasets, software, other artefacts) that the grant produced. Pre-GLS, each piece sat somewhere different. The funder lived in the Funder Registry (a Crossref-maintained list of funder organisations). The grant number was a free-text string in publisher metadata. The outputs were registered with DOIs at Crossref but without structured links back to the grant. The result was that the graph from funder through grant to output existed only in fragments.

    GLS provides the missing middle layer. A funder registers grants with Crossref via a dedicated deposit schema; each grant gets a DOI; the grant DOI carries structured metadata about the funder, the project, the awardees (with ORCID iDs where available), and the institutions (with ROR IDs). Publishers and other depositors then reference the grant DOI from the output’s metadata, closing the loop.

    2025 expansion: more funders, more metadata

    The major story of 2025 was funder participation. Through 2024 the GLS depositors were a small set of early adopters (Wellcome, the Templeton Foundation, ANR, a handful of others). 2025 added the major UK research councils (UKRI’s component councils now register grants via GLS), several EU H2020 and Horizon Europe streams (via OpenAIRE-mediated deposit), the Australian Research Council, and — significant for the US ecosystem — initial NSF and NIH pilots. NIH’s pilot is small (a few thousand R01 grants), but it signals direction.

    The metadata expansion was equally important. GLS 2025 added structured fields for: project abstract (free text but indexed), discipline classification (using Crossref-curated CODE FOR codes and OECD FoS), expected outputs, ethics-board identifiers, and — the field most relevant to CRediT integrators — a participants structure that names each grant participant with an ORCID iD and a role from a controlled vocabulary (principal investigator, co-investigator, collaborator, named researcher, fellow, named staff). This is not CRediT, but it interlocks with CRediT cleanly.

    The CRediT-GLS interlock

    Here is the integration pattern that is now possible. A grant is registered with GLS, getting a DOI and a structured participant list. A paper acknowledges the grant by including the grant DOI in its Crossref deposit. The paper’s JATS carries a CRediT contributor statement, which is also deposited to Crossref via the relationships block. ORCID consumes both deposits via the public API and can now answer: this researcher contributed to this paper in these CRediT roles; the paper acknowledges this grant; this researcher is on the grant participant list in this grant role.

    The query is structured and unambiguous. Pre-2025, it required string matching grant numbers and best-effort author-name disambiguation. Post-2025, the entire chain runs on PIDs. The CRediT adoption ledger at CASRAI tracks which publishers deposit CRediT to Crossref in the form that makes this work, and which still drop the qualifiers at the deposit step.

    What integrators need to do

    For publishers depositing content with Crossref, the 2025 GLS recommendations are: include grant DOIs in the funding section of every deposit where a grant is acknowledged; resolve and validate the grant DOI before deposit; carry CRediT roles with the degree-of-contribution qualifier in the contributor section. Crossref’s submission schema 5.4 supports all of this; older schema versions do not, and a number of publishers are still on 4.x.

    For institutional CRIS systems, the recommendation is to ingest grant DOIs into the funding record alongside the internal grant number, and to use the grant DOI as the join key when reconciling CRIS funding records against ORCID’s funding entries and Crossref’s content metadata. The CASRAI CRIS integration guide has been updated with the GLS ingestion patterns by major CRIS vendor.

    For funders not yet depositing to GLS, the question is what to do about historical grants. Crossref’s recommendation is to deposit prospectively (new awards from a chosen start date) and backfill historical grants over time. The funders that have done well at GLS uptake budgeted a small data-engineering effort over 6-12 months to backfill 5-10 years of historical grants from internal records.

    The RAiD-Crossref-GLS triangle

    An open question for 2026 is the relationship between GLS and RAiD. Both can identify a research project; both can carry participant, institution, funding, and output metadata; both have ISO-standard or de-facto-standard status. The honest answer is that they overlap meaningfully but serve different communities and emphases.

    GLS is funder-centric: a grant is the unit; the funder registers it; outputs reference it. RAiD is project-centric: a project can span multiple grants, multiple funders, multiple institutions, with the project itself the persistent unit. For a single-funder, single-project grant they are functionally identical. For a multi-funder collaboration (typical in clinical trials, large astronomy or particle physics projects, EU consortia), RAiD captures the project shape; the individual grants funding it would each be GLS-registered and the RAiD would reference them.

    The Crossref-DataCite-ARDC working group has begun work on a formal crosswalk that lets a GLS grant DOI declare a RAiD that it contributes to, and vice versa. This will not collapse the two but will let consumers traverse the graph in either direction.

    Consuming the data

    The Crossref REST API exposes GLS grants under /works with a type filter of grant; the relationship from a paper to its grant is in the paper’s relation block with relationship type is-funded-by. For ORCID-aware consumers, the ORCID 4.0 funding resource now carries the grant DOI as a primary identifier, with the Funder Registry entry as the funder.

    OpenAIRE consumes GLS deposits and exposes the resulting graph in its OpenAIRE Graph, which is the easiest single endpoint to query for the full funder-grant-output structure. For institutional consumers without the bandwidth to consume Crossref directly, OpenAIRE Graph is the recommended starting point.

    What’s still rough

    Two known limitations. First, the GLS participants structure does not yet carry CRediT roles directly; it carries grant participation roles, which are a coarser categorisation. This is by design — grant participation is not authorship — but it means that the question “which CRediT role did this person play on this grant” can only be answered indirectly, by intersecting grant participation with CRediT roles on the grant’s outputs. We expect this to be cleaned up in a future GLS schema revision.

    Second, historical-coverage gaps remain. Pre-2020 grants are almost entirely absent from GLS; 2020-2023 coverage is partial; 2024 onward is increasingly complete. Tools building on the GLS graph need to handle the missing-grant case gracefully.

    Related reading

  • Indigenous data governance: CARE Principles in practice

    The CARE Principles for Indigenous Data Governance were published in 2019 by the Global Indigenous Data Alliance (GIDA), expressing four principles – Collective benefit, Authority to control, Responsibility, Ethics – designed to sit alongside the FAIR principles when research data involves Indigenous Peoples, communities, lands, or knowledge. This post offers an introductory map of the CARE landscape in 2026, the relationships among the regional Indigenous data sovereignty movements that informed it, and the operational artefacts that researchers and institutions are using to apply CARE in practice. We write as outsiders to these traditions and rely on the published statements of Indigenous-led organisations; what follows is descriptive, not prescriptive, and any institution implementing CARE should engage directly with the communities whose data is in question.

    The CARE Principles

    The CARE Principles, drafted by Stephanie Russo Carroll, Maui Hudson, Tahu Kukutai, and colleagues through the GIDA, articulate that data governance is not only a question of technical FAIR-ness but of who has authority over data, who benefits, and what ethical commitments are owed. The four pillars are:

    • Collective benefit. Data ecosystems should be designed and function in ways that enable Indigenous Peoples to derive benefit from the data. Inclusive development and innovation; improved governance and citizen engagement; equitable outcomes.
    • Authority to control. Indigenous Peoples’ rights and interests in Indigenous data must be recognised and their authority to control such data must be empowered. Recognising rights and interests; data for governance; governance of data.
    • Responsibility. Those working with Indigenous data have a responsibility to share how those data are used. Capability for Indigenous communities; positive relationships; appropriate care for data.
    • Ethics. Indigenous Peoples’ rights and wellbeing should be the primary concern at all stages of the data life cycle. Minimising harm and maximising benefit; justice; future use.

    The principles are deliberately at the level of governance commitments, not the level of technical implementation. Their operationalisation depends on engagement with specific communities and their own governance institutions.

    The regional movements that preceded CARE

    CARE did not emerge in isolation. It is the international synthesis of regional Indigenous data-sovereignty movements that had been building governance frameworks for years.

    OCAP® Principles (Canada)

    The OCAP® Principles – Ownership, Control, Access, Possession – were articulated in 1998 by the First Nations Information Governance Centre (FNIGC) and have governed First Nations data in Canada since. OCAP is a registered trademark of FNIGC; the principles assert that First Nations have collective ownership of their information, control over how it is collected and used, access to it, and physical possession of it. FNIGC operates training programmes that researchers working with First Nations data are expected to complete; multiple Canadian Tri-Agency Indigenous research policies reference OCAP explicitly.

    Te Mana Raraunga (Aotearoa New Zealand)

    Te Mana Raraunga is the Māori Data Sovereignty Network, established 2015. Te Mana Raraunga articulates Māori data sovereignty rooted in tino rangatiratanga (self-determination) under Te Tiriti o Waitangi. The Network’s foundational statements include the 2018 Principles of Māori Data Sovereignty, which were among the documents informing CARE. The relationship between Te Mana Raraunga’s Māori-specific frame and the international CARE frame is one of mutual recognition; Te Mana Raraunga operates with the authority of Māori governance, not as an instance of an international standard.

    Maiam nayri Wingara (Australia)

    Maiam nayri Wingara, the Aboriginal and Torres Strait Islander Data Sovereignty Collective, was established in 2017 and articulated principles of Indigenous data sovereignty for Australia in 2018. The collective’s work emphasises the rights of Aboriginal and Torres Strait Islander peoples to control data about their people, communities, lands, and waters. The Australian Indigenous Health-Welfare Data Working Group and several federal agencies’ Indigenous data policies reference Maiam nayri Wingara’s frame.

    Other regional movements

    Indigenous data sovereignty movements with their own governance frameworks operate in many other contexts, including Sámi Council work in Sapmi, Native American data sovereignty organising in the United States (the United South and Eastern Tribes Tribal Health Program and others), Indigenous Latin American collectives, and others. The CARE Principles refer to and respect this plurality; they are not a substitute for any of these regional frameworks but a complement at international scale.

    How CARE relates to FAIR

    CARE and FAIR are designed to coexist. FAIR addresses technical interoperability and data reusability; CARE addresses governance authority and ethical commitments. A dataset can be both FAIR and CARE-compliant; a dataset can also be FAIR while failing CARE (technically open data that violates community authority); a dataset can be CARE-compliant while not openly FAIR (community-controlled data with restricted access in line with community decision).

    The GIDA’s published positioning is that CARE precedes FAIR when Indigenous data is involved: the questions of authority, benefit, responsibility, and ethics must be settled before the questions of findability, accessibility, interoperability, and reusability are operationalised. A FAIR-without-CARE approach to Indigenous data has historically reproduced harm; CARE asks researchers and institutions to do the governance work first.

    Free, Prior, and Informed Consent

    Free, Prior, and Informed Consent (FPIC) is the international human-rights principle, articulated in the UN Declaration on the Rights of Indigenous Peoples (UNDRIP, 2007) and widely adopted, that Indigenous Peoples must be consulted and consent obtained before any project affecting them, their lands, or their resources proceeds. FPIC applies to research projects involving Indigenous communities, knowledge, or data. The four elements – free (without coercion), prior (sufficiently in advance), informed (with adequate information), consent (with a community decision-making process) – are all substantive.

    FPIC operationalisation depends on the community in question. Some communities have formal protocols and Indigenous Research Ethics committees; others negotiate consent through community-leader engagement; others may decline participation. In all cases the timing of the consent process matters: FPIC sought after a project has been designed is generally not FPIC; FPIC must precede project design or at minimum precede any irreversible step.

    Traditional Knowledge Labels and Local Contexts

    Traditional Knowledge (TK) Labels and Biocultural (BC) Labels, developed by the Local Contexts initiative led by Jane Anderson and Kim Christen, are metadata labels that can be attached to datasets, archival records, or collection items to communicate community-defined permissions, attribution requirements, and cultural protocols. TK Labels include labels for attribution, non-commercial use, outreach, family or clan use, ceremonial use, and others; BC Labels cover biocultural specimens and data with similar granularity.

    The labels are not legal instruments by themselves; they are governance signals issued by communities that researchers and institutions are expected to respect. Several repositories (notably the Mukurtu CMS platform, also developed by Christen and colleagues) integrate TK and BC Labels natively. By 2026 several major museums, archives, and a small but growing number of institutional research repositories support TK Labels at the record level.

    Practical implementation for institutions

    An institution beginning to operationalise CARE alongside its FAIR practice would, in the broadest terms, attend to:

    1. Recognising the priority of community authority over data concerning Indigenous peoples, lands, and knowledge, and reflecting this in institutional research-data policy.
    2. Engaging with communities through their own governance institutions early, with FPIC understood as a substantive process not a checkbox.
    3. Adopting the relevant regional principles where applicable (OCAP in Canada, Te Mana Raraunga principles in Aotearoa, Maiam nayri Wingara in Australia, etc.) rather than treating CARE as a substitute.
    4. Supporting researchers in their institution with training, ethics-board capacity, and community-engagement resources; not pushing the burden onto Indigenous researchers within the institution.
    5. Implementing technical support for community-defined permissions (TK Labels, access-control models that respect community decision) in institutional repositories.
    6. Reporting transparently to communities about how data is used, with channels for community-initiated change to data status.

    Several institutional CRIS and repository vendors have begun adding CARE-aware functionality (TK Label support, community-attribution fields, access-control models that respect community-defined permissions). The CASRAI Indigenous data CARE domain tracks adoption.

    The integrity question

    The honest position for non-Indigenous researchers and institutions is that operationalising CARE well requires deferring to Indigenous-led governance, not designing one’s own “CARE-compliant” system. The literature is consistent on this point: the CARE Principles were developed by Indigenous-led organisations and their authoritative interpretation rests with those organisations and the communities they serve. The CARE Principles are not a checklist that an external institution can mark itself against and self-certify on.

    The implication for institutions and researchers is that the CARE work is relational and ongoing rather than one-time and administrative. The investment is in long-term partnerships with communities, capacity-building within Indigenous research leadership, and a willingness to share authority over how data flows into and out of institutional systems. The technical artefacts (TK Labels, FPIC processes, Mukurtu integrations) support the relational work; they do not substitute for it.

    Where to learn more

    For non-Indigenous researchers and institutions beginning this work, the foundational reading is the GIDA’s published statement of the CARE Principles, alongside the regional movements’ own foundational documents (FNIGC on OCAP, Te Mana Raraunga on Māori data sovereignty, Maiam nayri Wingara on Aboriginal and Torres Strait Islander data sovereignty). The Carroll, Hudson, Kukutai, et al. 2020 paper in Data Science Journal is the foundational scholarly reference for CARE. The Local Contexts initiative’s documentation is the foundational reference for TK and BC Labels. The Mukurtu CMS documentation is the foundational technical reference for community-controlled repository implementation.

    Related dictionary entries

    References

    Carroll, Hudson, Kukutai, et al., The CARE Principles for Indigenous Data Governance (Data Science Journal, 2020). GIDA, CARE Principles for Indigenous Data Governance (founding statement, 2019). First Nations Information Governance Centre, The First Nations Principles of OCAP® (foundational and ongoing publications). Te Mana Raraunga, Principles of Māori Data Sovereignty (2018). Maiam nayri Wingara, Indigenous Data Sovereignty Communique (2018). UN Declaration on the Rights of Indigenous Peoples (2007). Anderson and Christen, work on Traditional Knowledge Labels and the Local Contexts initiative (ongoing).

  • ORCID 4.0: the IDR roadmap and what it means for CASRAI integrations

    ORCID’s Integration and Data Roadmap (IDR) work, which culminated in late 2025 with the 4.0 release of the public and member APIs, is the most consequential PID infrastructure change of the year for anyone who cares about the contributor-affiliation-funding crosswalk. The headline is technical: a new contributions resource that supersedes the old works and employment pairing for representing what a researcher did, where, on whose money, and with whom. The implications reach into nearly every persistent-identifier integration CASRAI tracks.

    What 4.0 actually changes

    The pre-4.0 ORCID record was a federation of resource types: works (with DOIs), employment (with ROR organisation IDs), education, funding (with grant IDs and Funder Registry entries), peer reviews, and the like. Each was useful in isolation. None of them carried the relations between them in a structured form. If a researcher’s ORCID record listed a paper, an employment at the institution that hosted the work, and a grant that funded the work, those three facts sat in separate resources with no machine-readable link.

    4.0 introduces a top-level contribution entity that binds these. A contribution carries: a primary artefact (DOI, software identifier, dataset identifier, or RAiD), a set of CRediT roles with the degree-of-contribution qualifier, an affiliation in force at the time of the contribution (with ROR), funding in force at the time (with Funder Registry or ROR for the funder, plus the grant identifier and ideally a RAiD), and a temporal span. The relationships are explicit and queryable. A consuming system can ask: what did this researcher contribute, at this affiliation, under this grant, on this date? — and get an answer without inference.

    The CRediT-at-record-level integration matures

    The 2024 work to allow CRediT roles to live on an ORCID record (not just in publisher JATS) was the precursor to 4.0. The integration shipped, was widely adopted, and exposed two limitations that 4.0 closes. First, role assignments lived inside the work resource, making it awkward to express a Conceptualization role spanning several papers and datasets. Second, the qualifier was carried only at per-work granularity. 4.0 lets a CRediT role attach to a contribution that groups multiple artefacts, with the qualifier traveling with the contribution.

    Practical example: a researcher who is Lead for Conceptualization across a clinical trial’s primary paper, protocol paper, registered data, and statistical analysis plan should be representable that way. Pre-4.0, the assertion had to be repeated four times; post-4.0, it lives on the contribution entity. See the ORCID implementation guide for the API patterns.

    RAiD becomes a first-class citizen

    One of the unsung wins in 4.0 is the elevation of RAiD to a first-class identifier alongside DOI. Pre-4.0, RAiD could be carried in an ORCID funding resource as an external identifier, but the schema treated it as a second-tier metadata field. 4.0 adds RAiD to the primary identifier set for both contributions and funding, with the same validation and resolution support as DOI.

    This matters because RAiD is increasingly the canonical project-level identifier, and ORCID is increasingly the canonical person-level record. The interlock — researcher X contributed to project RAiD Y, which produced papers A, B, C — is now a structured query rather than a string-match exercise.

    Affiliation history with PIDs at both ends

    The 4.0 employment and affiliation model has been quietly tightened. Every affiliation now requires a ROR organisational ID at registration; legacy string-only affiliations are preserved but flagged. The optional department field accepts a ROR sub-organisation ID where one exists (the ROR hierarchy work has caught up to make this practical), or a free-text department name as a fallback. The result is that affiliation history on an ORCID record is now reliably machine-readable at the ROR ID level.

    For institutions running a CRIS, this closes a longstanding crosswalk gap. CRIS-to-ORCID deposit can now write structured affiliations that ORCID-to-CRIS retrieval can read back without ambiguity. The CASRAI CRIS integration guide has been updated with the 4.0 deposit patterns.

    What CASRAI integrations need to do

    Three things, in priority order.

    1. Update CRediT JATS round-trips. Publishers depositing structured CRediT to ORCID via the member API should switch to the contribution resource for new deposits. Legacy works-with-roles deposits will continue to be accepted through 2026 but will be migrated server-side in 2027. The CASRAI CRediT JATS integration patterns now include both the legacy and the 4.0 deposit forms; new integrators should implement only the 4.0 form.
    2. Validate ROR IDs at affiliation deposit. A CRIS or publisher pushing affiliation data to ORCID should resolve and validate the ROR ID before deposit. The 4.0 API will reject obviously bad ROR IDs at the schema layer but will accept ROR IDs that resolve to deprecated or merged records. A pre-deposit validation pass against the ROR public API catches the common error cases.
    3. Test the funding-to-contribution link. If your integration writes funding entries, link them explicitly to the contributions they funded via the new funded_by relation on the contribution resource. This is the integration point that was missing pre-4.0 and that downstream consumers (funder dashboards, institutional reporting) most want to query.

    Backwards compatibility and the migration window

    ORCID’s commitment is that the 3.x APIs remain available through end-of-2027, with the 4.0 API the recommended target from now. The data model migration is largely automatic for existing records: pre-existing works with associated employment and funding are projected into the contribution model server-side. Consumers reading via the 4.0 API will see contribution entities even for data that was deposited in the 3.x form.

    The one wrinkle is CRediT role assignments that were deposited in 3.x without explicit qualifiers. These project into the contribution model with no qualifier set — a valid state, but less informative than it could be. Publishers should re-deposit historical CRediT data with qualifiers where they have them during 2026.

    What this enables downstream

    The most interesting consequence of 4.0 is the ability to ask compound questions across the PID graph. Which researchers, affiliated with which institutions, contributed in which CRediT roles, to outputs funded by which funders, on which projects? — that query reduces to a structured traversal across ORCID, ROR, Crossref/DataCite, and the Funder Registry, with RAiD optionally tying the project layer together. The OpenAIRE Graph already operationalises a version of this; 4.0 makes it cleaner.

    For institutions, the practical implication is that reporting against funder mandates becomes substantially less manual. For publishers, the JATS-to-ORCID deposit becomes more valuable because it now persists in a queryable graph. For funders, the funder-PID-to-output traceability that ORCID has long promised starts to deliver at scale.

    What’s still missing

    4.0 does not solve everything. The contributor-affiliation-funding triple is now structured; the contributor-contributor relationship (collaboration graphs, mentorship) is not. A relationships resource is in development but not in 4.0. CARE-aligned identifiers for Indigenous researchers are also still in design.

    CASRAI’s integration tracking will follow 4.0 through 2026. The persistent-identifiers domain is being updated to reflect the contribution model; the ORCID federation page tracks member implementation.

    Related dictionary entries