CASRAI Dictionary

Tag: ARIA vs UKRI

  • REF 2029 Initial Decisions: What UKRI Confirmed and What’s Still Open for Consultation

    The REF 2029 initial decisions, published in 2023, confirmed decoupled outputs, HESA-based volume measures and no per-researcher output limits. On 10 December 2025, UKRI and Research England locked in revised element weightings and simplified guidance after an autumn pause – but the detailed Panel Criteria and Working Methods remain unpublished until Autumn 2026.

    The Research Excellence Framework (REF) is the UK’s system for assessing the quality and impact of research produced by higher education institutions, with REF 2029 the next exercise in the series and results due for publication in December 2029.

    What did the original REF 2029 Initial Decisions confirm in 2023?

    The four UK higher education funding bodies published Research Excellence Framework 2029: initial decisions and issues for further consultation (reference REF 2023/01) after reaching agreement through the Future Research Assessment Programme Board. This document set the high-level architecture that everything since has built on.

    Three assessment elements were renamed and reweighted to reflect a broader definition of research excellence: the outputs element became Contribution to Knowledge and Understanding, the environment element became People, Culture and Environment, and the impact element became Engagement and Impact. The funding bodies also confirmed that REF 2029 would move further away from assessing individuals, replacing staff-linked output counts with a volume measure drawn directly from Higher Education Statistics Agency (HESA) staff records.

    • Outputs decoupled from individual researchers under a “substantive link” policy tying outputs to the submitting institution instead.
    • No minimum or maximum number of outputs per individual, intended to widen inclusivity for early-career and returning researchers.
    • Unit-of-assessment structure retained largely as in REF 2021.
    • A short, targeted consultation on specific policy aspects launched immediately, alongside a discrete Open Access policy consultation.

    What changed when REF 2029 resumed after the 2025 pause?

    REF 2029 criteria-setting was paused in autumn 2025, following an announcement by UK Science Minister Lord Vallance that final guidance and Panel Criteria and Working Methods would not proceed to publication on the original schedule. On 10 December 2025, the REF team and the four funding bodies published updates confirming the exercise had resumed, shaped by further engagement with the sector and the expert REF panels.

    The People, Culture and Environment element was renamed again, to Strategy, People and Research Environment (SPRE), building on the REF 2021 Environment component and informed by the People, Culture and Environment Pilot report published the same day. Several simplifications were confirmed to reduce burden on institutions:

    • Contribution to Knowledge and Understanding (CKU) guidance simplified and clarified, with unit-level statements removed.
    • A recommended maximum of five outputs per researcher reinstated for clarity, while the minimum of one stays removed.
    • Impact case study requirements reduced to one for the smallest units, and the 2* qualifying threshold for underpinning research removed.
    • Limited portability introduced for long-form and extended-process research outputs, alongside simplified substantive-link requirements.

    Crucially, the funding bodies decided there would be no formal consultation on the guidance or the Panel Criteria and Working Methods, in order to protect the original timetable. REF 2029 panels began meeting in early 2026 to set criteria, and an Institutional-Level Working Group for SPRE is being established.

    How do the confirmed REF 2029 weightings compare at each stage?

    The element weightings moved between the 2023 Initial Decisions and the December 2025 update, reflecting sector feedback and the results of the PCE Pilot. Research England’s own framing is definitive: the funding bodies “refined these weightings having listened to the sector and considered the results of the PCE Pilot.”

    Element 2023 Initial Decisions weighting Confirmed 10 Dec 2025 weighting
    Strategy, People and Research Environment (formerly People, Culture and Environment) 25% 20%
    Contribution to Knowledge and Understanding 50% 55%
    Engagement and Impact 25% 25% (unchanged)

    REF underpins the allocation of roughly £2 billion annually in UK research funding, according to the REF team’s December 2025 announcement, which is why the funding bodies have prioritised timetable stability over a further round of formal consultation on these revised figures.

    Which REF 2029 questions remain open ahead of criteria publication?

    Despite the pace of confirmations through December 2025, the granular assessment criteria that panels and institutions ultimately submit against are not yet public. Per the REF 2029 timetable (last updated 10 December 2025), the sector is currently in the “onboarding of sub-panels” and “expert panels meet to develop guidance” phase covering winter, spring and summer 2026.

    • Panel Criteria and Working Methods – the unit-of-assessment-level detail institutions need for submission planning – is scheduled for Autumn 2026, not before.
    • The Institutional-Level Working Group for SPRE has not yet reported; its recommendations will shape how the 20%-weighted element is actually assessed.
    • The special requests process for exceptional submission circumstances launches only in the 2026-2027 winter window.
    • A long-form-output open access mandate was explicitly ruled out for REF 2029 itself but remains under discussion for the exercise that follows it.
    • Because the funding bodies opted out of formal consultation on the guidance and Panel Criteria and Working Methods, sector input on these final documents will run through panel and steering-group engagement rather than a published open call for responses.

    Answer-first Q&A on REF 2029

    What are the key changes for REF 2029?

    REF 2029 renames and reweights the three assessment elements, decouples outputs from named individuals under a substantive-link policy, replaces staff-linked output counts with a HESA-derived volume measure, and removes per-researcher output minimums while reinstating a recommended maximum of five outputs.

    What is the REF 2029 process?

    Institutions build submissions against Contribution to Knowledge and Understanding, Strategy, People and Research Environment, and Engagement and Impact, following Panel Criteria and Working Methods due in Autumn 2026. The submission window opens autumn 2027, closes autumn 2028, with results published December 2029.

    How many impact case studies are required for REF 2029?

    The December 2025 update reduced the impact case study requirement to a minimum of one for the smallest submitting units, retaining the REF 2021-style approach otherwise, and removed the 2* qualifying threshold previously applied to underpinning research.

    What publications are eligible for REF 2029?

    Outputs are eligible where they have a demonstrable substantive link to the submitting institution during the REF period, rather than being tied to a named individual’s employment dates, with simplified requirements and limited portability for long-form and extended-process research outputs.

    What the REF 2029 timeline means for institutions

    Research administrators cannot yet finalise submission strategy against unit-of-assessment-level criteria that do not exist in published form. What institutions can act on now are the settled structural decisions: the HESA-based volume measure, the substantive-link output policy, the five-output guideline, and the confirmed 20/55/25 weighting split.

    Because no formal consultation will precede the Autumn 2026 Final Guidance, the practical channel for institutional input is direct engagement with REF panels and the Institutional-Level Working Group for SPRE, not a published response process. This is a materially different sector-engagement model from the 2023-2024 Initial Decisions consultation and the 2024 Open Access consultation, both of which invited written responses.

    These structural questions sit alongside, but are distinct from, contributorship and authorship-attribution debates addressed by frameworks such as the authorship standards used elsewhere in scholarly communication, and the broader discipline of research administration that REF submission planning falls under.

    Outlook: REF 2029 heading toward Autumn 2026 Final Guidance

    REF 2029 has moved from broad principle to confirmed structure faster than its criteria have moved to publishable detail. The Initial Decisions set direction in 2023; the December 2025 update, following the 2025 pause, fixed weightings and simplified several requirements. What remains – the Panel Criteria and Working Methods – is the document institutions actually need to plan submissions, and it will not arrive before Autumn 2026. Until then, “REF 2029 initial decisions” describes a settled foundation, not a finished rulebook.

  • Inside ARIA: How It Funds Differently to UKRI

    The Advanced Research and Invention Agency (ARIA) is a UK non-departmental public body that funds high-risk, high-reward science through Programme Directors who design and run their own funding programmes, rather than through the peer-reviewed competitive grant calls used by UK Research and Innovation (UKRI). ARIA also holds a statutory exemption from Freedom of Information requests, a governance carve-out no UKRI council shares.

    ARIA is a British executive non-departmental public body, sponsored by the Department for Science, Innovation and Technology (DSIT), established on 26 January 2023 under the Advanced Research and Invention Agency Act 2022. It operates independently of UKRI, with its own statutory ten-year mandate to fund research that is “risky, uncertain and speculative in nature.”

    What is ARIA and who runs it?

    ARIA was proposed by Dominic Cummings as a UK equivalent to the US Defense Advanced Research Projects Agency (DARPA), and it received Royal Assent as the Advanced Research and Invention Agency Act 2022. It launched with an initial £800 million in government funding, and the June 2025 Spending Review committed a further minimum of £1 billion over the 2025–2029 period.

    ARIA is small by design. Its FY2024/25 annual report and accounts record just 53 employees and a £27.6 million annual budget. Dr Kathleen Fisher became ARIA’s CEO in February 2026, succeeding founding CEO Ilan Gur, who announced his departure in June 2025. Matt Clifford CBE remains Chair. The board also includes the Government Chief Scientific Adviser and advisors such as Sir Demis Hassabis of Google DeepMind.

    How does ARIA’s Programme Director model actually work?

    Rather than issuing standing competitive calls, ARIA gives named Programme Directors — senior scientists and engineers hired for fixed terms — the authority to define an “opportunity space” and then build and fund a multi-year programme inside it. This is the single biggest structural difference from UKRI’s council-based, panel-reviewed system.

    ARIA’s own published model breaks funding into four channels:

    • Opportunity spaces — broad areas judged highly consequential, under-explored, and ready for new investment.
    • Programmes — multi-year, multi-institution R&D efforts built inside an opportunity space by a Programme Director.
    • Opportunity seeds — smaller, less structured awards to individuals or teams pursuing aligned but unprogrammed ideas.
    • Activation Partners — non-profit and commercial partners who supply entrepreneurial talent and capital to help funded research reach deployment.

    ARIA’s first cohort of eight Programme Directors joined in October 2023, with a second cohort of eight following in April 2025. Notable live programmes include Forecasting Tipping Points (a 27-team, £81 million climate-monitoring effort led by Sarah Bohndiek and Gemma Bale), Scaling Compute (targeting a 1,000-fold reduction in AI compute costs, led by Suraj Bramhavar), and a £50 million outdoor solar geoengineering research programme.

    Funding terms are also structurally different. ARIA does not retain intellectual property rights over funded work, generally does not require match funding, and does not take equity itself, though it caps the equity that funded organisations can hold in spinouts.

    ARIA vs UKRI: how the two funding models compare

    UKRI remains the UK’s primary public research funder, distributing a combined budget of several billion pounds a year across seven research councils, Innovate UK and Research England through competitive, peer-reviewed calls. ARIA is built to complement, not replace, that system by taking on the speculative, early-stage bets UKRI’s accountability structures are not designed to absorb.

    Dimension UKRI ARIA
    Funding decision-maker Peer-review panels, per council Individual Programme Directors
    Primary mechanism Competitive, published grant calls Programme Director-defined “opportunity spaces”
    Risk tolerance Assessed feasibility, incremental risk Statutory mandate to fund “risky, uncertain and speculative” work
    IP treatment Varies by scheme ARIA does not retain IP rights
    FOI status Subject to Freedom of Information Act 2000 Statutorily exempt from FOI requests
    Legal basis Higher Education and Research Act 2017 Advanced Research and Invention Agency Act 2022

    Institutions engaging with ARIA also encounter a different bidding rhythm: rather than responding to fixed annual calls, Creators (ARIA’s term for funded teams, ranging from individual PhD researchers to large organisations) are often recruited directly into a Programme Director’s opportunity space, or apply to time-limited programme-specific calls announced on a rolling basis.

    Why is ARIA exempt from FOI, and what does that mean for accountability?

    The Advanced Research and Invention Agency Act 2022 gave ARIA a statutory exemption from the Freedom of Information Act 2000, a provision Cummings pushed for on the grounds that FOI slows fast, high-risk decision-making. Labour MPs, including Dawn Butler, later attempted to repeal the exemption in Parliament; the attempt was defeated, and the exemption stands.

    The exemption does not remove all scrutiny: ARIA still publishes annual reports and accounts, and its executives have appeared before the Commons Science, Innovation and Technology Committee, where CEO Ilan Gur confirmed in September 2025 evidence that UKRI’s Economic and Social Research Council has no authority to halt ARIA’s projects. In May 2026, this reduced-transparency model drew scrutiny after The Guardian reported that ARIA had allocated roughly £23 million to nine US-based technology companies and £29.4 million to three US venture capital groups, prompting economist Cecilia Rikap to question whether public money was expanding US tech-sector capacity rather than UK capability — a debate that reduced FOI visibility makes harder for outside researchers to independently audit.

    What does this mean for institutions bidding into ARIA’s 2026 calls?

    For research administrators and institutional leaders, the practical implication is a fundamentally different due-diligence process. Where UKRI bids are built around published assessment criteria and reviewer panels, ARIA engagement typically starts with a relationship to a specific Programme Director’s opportunity space, and success depends on fit with that programme’s technical thesis rather than a scored proposal against generic criteria.

    Institutions should also budget for ARIA’s lighter contractual overhead — no mandatory match funding and no IP retention lowers the administrative burden compared with many UKRI schemes — while recognising that ARIA’s rolling, programme-specific calls require closer, more continuous horizon-scanning of ARIA’s opportunity spaces than the UKRI funding calendar demands. With a confirmed minimum £1 billion available over 2025–2029 and an expanding programme slate under incoming CEO Kathleen Fisher, ARIA calls are likely to grow in frequency and disciplinary breadth through 2026.

    Answer-first Q&A

    Is ARIA a government agency?

    ARIA is an executive non-departmental public body, not a government department. It is sponsored by the Department for Science, Innovation and Technology (DSIT) and funded through Parliament, but the Advanced Research and Invention Agency Act 2022 gives it statutory independence from day-to-day ministerial direction over individual funding decisions.

    Is ARIA part of UKRI?

    No. ARIA is a separate legal body from UK Research and Innovation, created by its own Act of Parliament rather than sitting inside UKRI’s research-council structure. The two organisations are designed to be complementary, with ARIA funding higher-risk work that UKRI’s peer-review processes are not structured to support.

    Who is the CEO of ARIA?

    Dr Kathleen Fisher has been ARIA’s CEO since February 2026, after being named CEO Elect in November 2025. She succeeded founding CEO Ilan Gur, a former ARPA-E Programme Director who announced his departure in June 2025. Matt Clifford CBE has chaired ARIA’s board since 2022.

    Is there a UK version of NASA?

    Not via ARIA. ARIA is modelled on DARPA, the US Department of Defense’s high-risk research agency, not NASA, the US space agency. The UK’s space activities sit instead with the UK Space Agency, a separate executive agency also sponsored by DSIT.

    Looking ahead

    ARIA’s structural bet — Programme Director autonomy, reduced FOI accountability, and tolerance for failed projects — is now backed by a confirmed £1 billion funding floor through 2029, making it a permanent rather than experimental feature of the UK funding landscape. For institutions and research administration teams weighing whether to pursue an ARIA opportunity space or a UKRI call, the choice increasingly comes down to risk appetite and governance tolerance rather than funding availability alone.