CASRAI Dictionary

Category: Policy & Funding News

Reporting and briefings on external policy, regulatory, and funder developments affecting the research community worldwide.

  • AI Safety Institute Network: How Cross-Border Testing Works

    The AI safety institute network is a group of ten government-backed bodies — currently Australia, Canada, the EU, France, Japan, Kenya, the Republic of Korea, Singapore, the UK and the US — that coordinate technical methods for evaluating advanced AI models before and after deployment. It is not one institute; it is a forum for shared testing standards, joint exercises and information-sharing between national evaluators, formally renamed in February 2026 to reflect a wider science-of-evaluation mandate.

    The AI safety institute network traces to the May 2024 Seoul AI Summit, where the UK, US, Japan, Singapore, the EU and several other governments agreed to establish an international forum for AI safety science. Since then it has run joint model-testing pilots, absorbed a new member (Kenya), and changed its own name — developments that most explainer content published in 2024–25 has not caught up with.

    What is the AI safety institute network?

    The network is a technical forum, not a regulator. Each participating jurisdiction runs its own domestic AI safety or security institute — a government office, typically sitting inside a science or digital ministry — and the network is the channel through which those offices align methods rather than rules.

    As of February 2026 its formal name is the International Network for Advanced AI Measurement, Evaluation and Science, having operated since its November 2024 founding as the “International Network of AI Safety Institutes.” The rename, announced on the UK AI Security Institute’s blog, signals a shift from a safety-branding forum toward a standing scientific body focused on measurement methodology.

    • Established: November 2024, following the Seoul Statement of Intent signed at the May 2024 AI Seoul Summit
    • Current members: Australia, Canada, the European Union, France, Japan, Kenya, the Republic of Korea, Singapore, the United Kingdom and the United States
    • 2026 Network Coordinator: the United Kingdom, via the AI Security Institute (AISI)
    • Mandate: build internationally recognised methods for measuring and evaluating advanced AI capabilities and risks

    How does cross-border AI model testing actually work?

    Cross-border testing runs on bilateral agreements and joint exercises rather than a single binding treaty. The clearest example is the UK–US Memorandum of Understanding signed on 1 April 2024, under which the two institutes committed to a shared approach to model evaluations, at least one joint test of a publicly accessible model, technical-research collaboration and staff exchanges.

    Member institutes have already run several joint exercises. A pilot evaluation of an open foundation model tested interoperability between two toolkits — the UK’s open-source Inspect framework and Singapore’s Moonshot platform — checking whether results from different infrastructure could be compared directly.

    • A multilingual evaluation led jointly by Singapore, Japan and the UK translated and validated benchmark questions into ten languages to test privacy, crime, intellectual-property and jailbreak-resistance risks across languages
    • A joint exercise on “agentic” AI systems examined risks including sensitive-information leakage, fraud and cybersecurity in models capable of autonomous planning and task execution
    • Information-sharing covers model limitations, capabilities and documented safety incidents, subject to each institute’s domestic law

    No member institute has ceded sovereign testing authority to the network. Each government still decides what to evaluate and publish; the network’s value is methodological convergence, not enforcement.

    What changed when the network renamed itself in 2026?

    At a technical meeting held alongside the NeurIPS conference in San Diego in late 2025, members reported growing consensus on evaluation principles and surfaced open methodological questions in a February 2026 blog post — evidence the network has moved from launch-phase diplomacy to a working scientific programme.

    Consensus areas reported include: evaluations must state clear objectives before they run; results must be transparent and reproducible; testing parameters should match the evaluation’s purpose (best-case elicitation versus realistic use); and reports should separate individual tests and stay legible to non-specialist policymakers.

    Unresolved questions include how much methodological detail evaluators can publish without enabling replication of dangerous capabilities, how flexible reporting templates should be, and how to evaluate deployed AI systems — not just underlying models. The network’s next meeting is at the India AI Impact Summit, where members plan to test approaches against real-world use cases.

    How do the national institutes compare?

    Funding, remit and regulatory posture vary sharply between members — “the AI safety institute” is a misleading singular, since at least six materially different institutional models operate under one network.

    Jurisdiction Institute Parent body Reported funding Primary posture
    United Kingdom AI Security Institute (AISI, formerly AI Safety Institute) Department for Science, Innovation and Technology £100 million secured until 2030 Voluntary evaluations, R&D, 2026 Network Coordinator
    United States Center for AI Standards and Innovation (CAISI, formerly US AISI) NIST $10 million for FY2024/25 Voluntary evaluations and standards R&D
    European Union EU AI Office DG Connect €46.5 million setup funding Binding regulation — enforces the EU AI Act
    Singapore Digital Trust Centre / AI Verify Foundation IMDA $37 million setup funding Testing tooling and content-assurance standards
    Japan Japan AI Safety Institute IPA Not separately disclosed Evaluation methodology and standards research
    Canada Canadian AI Safety Institute ISED Not separately disclosed R&D and international coordination

    The EU is the outlier: its AI Office is the only member with statutory enforcement power, administering the EU AI Act’s obligations on providers of general-purpose models with systemic risk. Every other member’s work is voluntary — which is why joint testing protocols, not legislation, are the network’s main coordination tool.

    Where can academic researchers plug into pre-deployment evaluation work?

    Pre-deployment evaluation is not a closed government function. Several member institutes have built explicit, funded routes for academic researchers rather than treating testing as an internal civil-service exercise.

    • Open evaluation tooling: the UK AISI’s Inspect framework is published as open-source software, so university research groups can build and run their own model evaluations using the same infrastructure member institutes use, rather than reverse-engineering proprietary test harnesses
    • Grant-funded safety research: AISI’s Systemic Safety Grants programme has funded academic work on systemic AI risks, alongside The Alignment Project — a multi-funder grants and compute programme backed by AISI, UKRI, Schmidt Sciences and several frontier AI developers supporting alignment research beyond government labs
    • Conference-linked workshops: the network’s 2025 technical meeting was held alongside NeurIPS to exchange methods with academic and industry researchers, not as a closed intergovernmental session
    • Multilingual and cross-cultural evaluation gaps: the network has flagged that evaluation science for non-English, non-Western contexts is underdeveloped — a direct, citable entry point for university linguistics and computational social science groups

    For institutional leaders, the practical implication is that AI safety evaluation now sits alongside conventional funder compliance work: grant terms from bodies co-funding alignment research carry reporting and reproducibility expectations that research administration teams already apply to other externally funded programmes.

    Answer-first questions on the AI safety institute network

    Which countries have AI safety institutes?

    Ten members currently participate in the International Network for Advanced AI Measurement, Evaluation and Science: Australia, Canada, the European Union, France, Japan, Kenya, the Republic of Korea, Singapore, the United Kingdom and the United States, each running its own domestic evaluation body.

    Did the AI Safety Institute change its name?

    Yes. In 2025 the UK’s AI Safety Institute was renamed the AI Security Institute, and its US counterpart became the Center for AI Standards and Innovation (CAISI). In February 2026 the international network itself was renamed from the “International Network of AI Safety Institutes” to the “International Network for Advanced AI Measurement, Evaluation and Science.”

    Are the network’s testing protocols legally binding?

    No single binding treaty governs the whole network. Cooperation runs through bilateral instruments such as the UK–US Memorandum of Understanding and voluntary joint exercises; the EU AI Office is the sole member with statutory enforcement power, via the EU AI Act.

    Who coordinates the network in 2026?

    The United Kingdom, through its AI Security Institute, holds the 2026 Network Coordinator role and has committed to converting the group’s evaluation-science consensus into detailed best-practice documentation before the network’s next meeting at the India AI Impact Summit.

    What happens next

    The network’s trajectory — from a 2024 diplomatic statement to a 2026 renamed scientific body with a rotating coordinator and a published consensus-and-open-questions agenda — suggests consolidation around evaluation methodology rather than a drift into national silos. Kenya’s addition also signals a deliberate widening beyond the original G7-plus-Singapore core.

    For universities, funders and publishers tracking AI governance, the signal to watch is the best-practice documentation the UK has committed to during its 2026 coordinator term, and whether evaluating deployed AI systems — not just underlying models — gets resolved before the India AI Impact Summit.

  • UK AI Regulatory Framework: EU Sandboxes to 2027

    The UK AI regulatory framework relies on existing sector regulators and five cross-sectoral principles rather than a single AI law, while a related EU milestone has just slipped: Article 57 of the EU AI Act required every member state to launch a national AI regulatory sandbox by 2 August 2026, and the EU’s Digital Omnibus simplification package has now pushed that deadline to 2 August 2027. For research institutions piloting AI in admissions, exam proctoring, or research-assistant tools, the delay changes when a supervised testing route becomes available — and it puts a spotlight on what the UK offers instead.

    An AI regulatory sandbox is a supervised legal and technical environment, established by a national competent authority, in which providers can develop, test, and validate innovative AI systems under direct regulatory oversight before those systems are placed on the market.

    What is an AI regulatory sandbox under Article 57?

    Article 57 of Regulation (EU) 2024/1689 — the EU AI Act — requires each member state to ensure its competent authorities establish at least one national AI regulatory sandbox. Inside the sandbox, providers develop, train, validate, and test AI systems under a supervised programme agreed with the regulator, with derogations available for limited real-world testing before a product goes to market.

    The mechanism exists because conformity assessment for high-risk AI systems is otherwise a one-shot, post-hoc exercise. A sandbox lets a university, a health authority, or a fintech firm iterate on a system’s design with a regulator in the room, reducing the risk of building a product that fails assessment after deployment. The AI Act entered into force on 1 August 2024 and becomes fully applicable on 2 August 2026, with obligations phased in across that period.

    Why did the 2026 sandbox deadline slip?

    The original Article 57 deadline required sandboxes to be operational by 2 August 2026 — the same date the AI Act’s general obligations take full effect. By early 2026, the European Parliament’s own think tank was reporting that the European Commission had not yet adopted the implementing act setting out common rules for how sandboxes should operate, leaving member states without the technical detail needed to stand theirs up on schedule.

    Several factors compounded the delay:

    • No implementing act: member states lacked Commission guidance on common sandbox rules until late in the schedule.
    • Resourcing: newly designated national AI authorities lacked the staff and budget sandboxes require.
    • Sequencing: sandboxes matter most for high-risk systems, and those detailed obligations do not apply until August 2027 anyway.

    What does the Digital Omnibus actually change?

    The Digital Omnibus is the European Commission’s 2026 simplification package for digital-rules legislation, including targeted amendments to AI Act deadlines. Under the package, the deadline for national AI regulatory sandboxes moves from 2 August 2026 to 2 August 2027 — aligning it with the date the Act’s detailed high-risk system obligations become enforceable, rather than with the earlier general-applicability date.

    The table below sets out how the EU timeline compares with the sandbox-equivalent mechanisms already running in the UK, which is not an EU member state and is not directly bound by Article 57.

    Mechanism Jurisdiction Legal basis Status / deadline
    National AI regulatory sandbox Each EU member state AI Act Article 57 (Regulation (EU) 2024/1689) Delayed from 2 Aug 2026 to 2 Aug 2027 under the Digital Omnibus
    FCA Regulatory Sandbox UK, financial services FCA innovation framework Running in cohorts since 2016
    ICO Regulatory Sandbox UK, data protection ICO service, independent of the AI Act Ongoing, rolling applications
    AI Growth Labs UK, cross-sector Follows the AI Opportunities Action Plan Pilot phase, sector-by-sector rollout

    Does the UK AI regulatory framework offer an equivalent?

    The UK AI regulatory framework is a pro-innovation, context-specific model set out in the 2023 white paper “AI regulation: a pro-innovation approach”. Instead of a horizontal AI statute, existing regulators — the Information Commissioner’s Office (ICO), the Competition and Markets Authority (CMA), and the Financial Conduct Authority (FCA) among them — apply five cross-sectoral principles: safety, security and robustness; appropriate transparency and explainability; fairness; accountability and governance; and contestability and redress.

    The UK has no Article 57 equivalent written into statute, but it is not starting from zero. The FCA has run a financial-services regulatory sandbox since 2016, the ICO already operates its own sandbox for organisations testing innovative, personal-data-driven products, and the government’s newer AI Growth Labs initiative is designed to pilot AI applications that existing rules would otherwise slow down. The gap is horizontal, cross-sector coverage of the kind Article 57 mandates for AI specifically — which matters because the AI Act’s extraterritorial scope catches any provider or deployer placing an AI system on the EU market or serving EU-based users, including UK universities with EU campuses, Erasmus partnerships, or platforms used by EU-resident students and researchers.

    What should research institutions piloting AI do now?

    Three categories of university AI pilot sit closest to this regulatory activity, and two of them are explicitly named as high-risk under Annex III of the AI Act: systems used to evaluate learning outcomes or assign students to institutions (admissions algorithms), and systems used to monitor or detect prohibited behaviour during tests (exam proctoring). Research-assistant models are not automatically high-risk but can trigger obligations depending on how their outputs are used in decision-making.

    Practical steps institutions can take while sandbox access is delayed:

    • Map pilots against Annex III now, since admissions and proctoring tools carry the highest compliance burden once high-risk obligations bite.
    • Use available UK sandboxes — the ICO’s service in particular — for pilots with a significant personal-data component, since that route does not depend on the EU timeline.
    • Track sandbox announcements in EU jurisdictions where the institution has a legal presence, so an application can be lodged as soon as one opens.
    • Document testing activity conducted before 2027; sandbox participation typically requires evidence of a structured development process, not a blank pilot history.

    Answer-first questions on sandboxes and the delay

    What is the deadline for AI regulatory sandboxes now?

    Under the Digital Omnibus, EU member states must have at least one operational national AI regulatory sandbox by 2 August 2027, one year later than the original Article 57 deadline of 2 August 2026. The new date aligns with when the AI Act’s detailed high-risk obligations take full effect.

    Which EU countries missed the original 2026 sandbox deadline?

    By the original deadline, most member states had not launched an operational sandbox, largely because the European Commission’s implementing act setting common sandbox rules had not been adopted in time. Newly designated national AI authorities also lacked the staffing to meet the schedule unassisted.

    Does the UK have to comply with the EU AI Act?

    The UK is not an EU member state, so Article 57 does not bind it directly. However, the AI Act’s extraterritorial scope applies to any provider or deployer placing an AI system on the EU market or serving EU-based users — a live issue for UK universities with EU partnerships or EU-resident users.

    Are university admissions and proctoring tools classified as high-risk AI?

    Annex III of the EU AI Act explicitly lists AI systems used for admission or assignment to educational institutions, and for monitoring or detecting prohibited student behaviour during tests, as high-risk applications. Both categories face the Act’s strictest conformity, documentation, and human-oversight requirements.

    Outlook: what comes next

    The sandbox delay buys implementers time, but it does not change the substance of what Article 57 sandboxes are for or which university AI pilots will eventually need them. Institutions that map their admissions, proctoring, and research-assistant pilots against Annex III now — and use existing UK routes such as the ICO sandbox in the interim — will be positioned to apply the moment national EU sandboxes open in 2027, rather than starting that process from scratch.

    Research administrators coordinating these pilots across institutional and cross-border governance structures may find it useful to review how research administration functions are adapting their compliance workflows to AI-specific regulatory requirements more broadly.

  • Digital Omnibus AI Act: New 2027 Deadlines

    The Digital Omnibus AI Act agreement, reached by EU co-legislators on 7 May 2026, postpones the AI Act’s high-risk obligations to 2 December 2027 for standalone systems and 2 August 2028 for product-embedded systems, pushes the national AI regulatory sandbox deadline from 2 August 2026 to 2 August 2027, and shortens the AI-generated-content labelling grace period to a new deadline of 2 December 2026. Prohibited-practice and general-purpose-AI (GPAI) obligations already in force are unaffected.

    The Digital Omnibus on AI is the EU’s amending regulation to Regulation (EU) 2024/1689 (the AI Act) that recalibrates several implementation deadlines and simplifies selected compliance requirements without altering the Act’s underlying risk-based framework.

    What Has Changed Under the Digital Omnibus?

    The European Commission published its Digital Omnibus on AI proposal on 19 November 2025, and the Council presidency and European Parliament negotiators reached a provisional political agreement on 7 May 2026. The European Parliament granted final approval on 16 June 2026. As of early July 2026, formal Council adoption and publication in the Official Journal are still pending, with completion expected by 2 August 2026 — the very date the original high-risk deadline would otherwise have taken effect.

    Until the amending regulation is published, the AI Act’s original text remains binding law. This is a narrow but real compliance-planning window: institutions cannot yet treat the new dates as legally settled, only as highly likely.

    The package also adds a new prohibition on AI systems that generate child sexual abuse material (CSAM) or non-consensual sexually explicit content (“nudifier” apps), reinstates the EU database registration requirement for AI systems exempted from high-risk classification, and reverts a proposed relaxation on processing special category data for bias detection back to a strict-necessity test.

    What Are the New AI Act Compliance Deadlines?

    The revised timeline replaces the Commission’s original “standards-linked” mechanism with fixed calendar dates, giving institutions a firm planning horizon rather than a moving target tied to standardisation progress.

    Obligation Original deadline New deadline Change
    Standalone high-risk systems (Annex III: education access, employment/HR, credit scoring, critical infrastructure, law enforcement) 2 August 2026 2 December 2027 16-month delay
    High-risk systems embedded in regulated products (Annex I: medical devices, machinery, toys) 2 August 2027 2 August 2028 12-month delay
    AI-generated content labelling/watermarking (Article 50(2)) 2 August 2026 2 December 2026 Grace period cut to 4 months
    National AI regulatory sandbox establishment (Article 57) 2 August 2026 2 August 2027 12-month delay
    Ban on CSAM/non-consensual intimate AI content Not previously prohibited 2 December 2026 New obligation
    Prohibited AI practices (Article 5) 2 February 2025 Unchanged Already in force
    GPAI model obligations (Articles 53–55) 2 August 2025 Unchanged Already in force

    Per the Council’s 7 May 2026 press release, “the provisional agreement also introduces a fixed timeline for the delayed application of high-risk rules: the new application dates would be 2 December 2027 for stand-alone high-risk AI systems and 2 August 2028 for high-risk AI systems embedded in products.” The same text confirms the sandbox deadline is postponed “until 2 August 2027.”

    Which AI Act Obligations Still Apply in 2026?

    Despite the headline delays, several obligations remain live this year. Institutions should not read “Digital Omnibus” as “AI Act paused.” The Act’s prohibited-practice regime and its general-purpose-AI rules were untouched by the negotiations.

    • Prohibited AI practices under Article 5 (e.g. social scoring, certain biometric categorisation, manipulative systems) have applied since 2 February 2025 and remain fully enforceable.
    • GPAI model provider obligations (transparency documentation, copyright-policy summaries, systemic-risk assessment for the most capable models) have applied since 2 August 2025.
    • Most Article 50 transparency duties — informing individuals they are interacting with an AI system, or that content is AI-generated — still take effect from 2 August 2026; only the specific machine-readable watermarking sub-obligation is delayed to 2 December 2026.
    • The narrowed research exemption in Article 2(6)/(8) is unchanged: it still covers only AI systems developed for the “sole purpose” of scientific research and development, and does not extend to real-world testing outside that narrow scope — a gap industry and legal commentators flagged but the Omnibus did not close.

    What Should Research Institutions Do Now?

    The Annex III high-risk categories map directly onto functions many universities, funders, and research offices already run or procure: “access to education and vocational training,” and “employment-related uses” covering recruitment, performance monitoring, and promotion decisions. Any admissions-scoring tool, proctoring system, or HR-screening AI a research institution uses now has until 2 December 2027 rather than August 2026 to meet high-risk documentation, human-oversight, and conformity-assessment requirements.

    That extra runway does not extend to everything an institution touches:

    • GPAI-based research tools (foundation models used in text/data mining, literature synthesis, or research-assistant products) are already subject to provider transparency obligations since August 2025 — this was not delayed and should already be reflected in procurement due diligence.
    • AI regulatory sandboxes, a route some national research funders and public research bodies planned to use for supervised testing of experimental AI tools, will not be mandatory at national level until 2 August 2027 — a year later than institutions may have budgeted for.
    • The research exemption remains narrow. Institutions running real-world pilots of AI tools (learning-analytics trials, clinical-AI validation studies) outside a controlled research-only environment should not assume blanket exemption; the classification tests apply as originally drafted.
    • AI-content labelling (Article 50(2), now due 2 December 2026) is directly relevant to scholarly publishing workflows: journals, repositories, and preprint servers using generative tools in editorial or production processes should track this date alongside their existing disclosure policies for AI-assisted content.

    Research administration offices coordinating compliance calendars should treat 2 December 2027 and 2 August 2028 as the two hard deadlines for high-risk systems, while keeping the unaffected 2025-dated GPAI and prohibited-practice obligations on their existing tracker — the Digital Omnibus changes the pace of the high-risk regime, not its scope.

    Answer-First Q&A

    What is the timeframe for the AI Act?

    The AI Act entered into force on 1 August 2024. Prohibited practices applied from 2 February 2025 and GPAI obligations from 2 August 2025. Following the Digital Omnibus, standalone high-risk systems now apply from 2 December 2027 and product-embedded high-risk systems from 2 August 2028.

    When do the AI Act’s high-risk obligations now apply?

    Under the provisional agreement, standalone Annex III high-risk systems (education, employment, credit, critical infrastructure) must comply by 2 December 2027. Annex I product-embedded systems (medical devices, machinery) have until 2 August 2028 — 16 and 12 months later than the AI Act’s original dates, respectively.

    Does the Digital Omnibus delay the AI Act sandbox deadline?

    Yes. The national AI regulatory sandbox deadline under Article 57 moves from 2 August 2026 to 2 August 2027, giving competent authorities an extra year to build supervised testing environments for innovators and public bodies.

    What AI Act obligations still apply in 2026?

    Prohibited practices and GPAI model obligations remain fully in force, having applied since 2025. Most Article 50 transparency duties still take effect on 2 August 2026, and the new CSAM/nudifier ban and AI-content watermarking sub-obligation both land on 2 December 2026.

    What Happens Next?

    The amending regulation still requires formal Council adoption and publication in the Official Journal before the new dates become legally binding, a process both the Council and independent legal analysis expect to conclude by 2 August 2026. Research institutions should build compliance calendars around the dates above now, while monitoring the Official Journal publication to confirm the fixed timeline takes definitive legal effect, and continue tracking CEN-CENELEC’s harmonised AI standards, whose slower-than-expected delivery was the stated driver for the entire postponement.

  • UK Research Funding Volatility 2026: Key Signals

    UK research funding volatility 2026 refers to the disruption research offices are managing this year as UK Research and Innovation (UKRI) replaces its council-by-council allocation model with a new three-part “buckets” structure, pauses several applicant-led grant routes, and holds discretionary research funding flat in cash terms while overall spending rises. For research administrators, the practical effect is a period of scheme-by-scheme uncertainty layered on top of institutional financial strain.

    Research funding volatility, in this context, is the combination of budget reallocation, temporary application pauses, and system migration that makes it harder for a research office to predict which schemes will be open, on what terms, and on what platform, in any given month of 2026.

    Contents

    Why Is UK Research Funding Volatile in 2026?

    UK research funding volatility in 2026 stems from three overlapping pressures hitting research offices simultaneously. UKRI is mid-way through a multi-year restructuring of how it allocates money between now and 2030. Several research councils have paused or reopened applicant-led grant routes within months of each other. And university finances are already stretched, so any funder-side disruption lands on institutions with little slack to absorb it.

    The Russell Group reports that 45% of English higher education providers are forecasting a deficit for the 2025/26 financial year, with universities in Wales, Scotland and Northern Ireland facing even more acute pressure. Against that backdrop, a paused grant call or a delayed platform migration is not a minor administrative inconvenience — it is a cash-flow and workforce-planning risk for the office managing it.

    How Is UKRI Restructuring Its Funding Model?

    From April 2026, UKRI has organised its budget into three funding “buckets”: curiosity-driven research, strategic government and societal priorities, and support for innovative companies. This replaces the previous council-led allocation logic with a cross-cutting structure intended to make funding priorities more explicit and outcome-focused.

    UKRI’s own December 2025 budget-allocations explainer confirms that the organisation’s total research and innovation budget is on a path to rise towards £10 billion a year by 2030, even as the mix shifts. Sector analysis published by Wonkhe on the CEO’s budget mapping, and coverage by the Institute of Physics, both describe discretionary curiosity-driven funding as held flat in cash terms over the same period — a real-terms reduction once inflation is applied, even as strategic-priority and innovation-linked funding lines grow.

    • Curiosity-driven research: open, investigator-led schemes; budget held broadly flat in cash terms into the near future.
    • Strategic priorities: government- and society-aligned mission funding, expected to absorb a growing share of the uplift.
    • Innovative companies: commercialisation and translation funding routed largely through Innovate UK.

    UKRI’s Chief Executive, Professor Sir Ian Chapman, confirmed to the House of Commons Science, Innovation and Technology Committee in March 2026 that the organisation will publish a new single delivery plan for the 2026/27 financial year, intended to give the sector its first detailed, comparable view of spending under the new bucket model.

    Which Funding Pauses and Cuts Have Hit Research Offices?

    Alongside the structural changes, individual councils have paused or reduced specific routes. According to analysis by the Campaign for Science and Engineering (CaSE), published in February 2026, three research councils — the Medical Research Council (MRC), the Biotechnology and Biological Sciences Research Council (BBSRC) and the Engineering and Physical Sciences Research Council (EPSRC) — paused applicant-led grant routes in the same period, while the Science and Technology Facilities Council (STFC) began cost reductions driven by rising international subscription costs.

    UKRI’s own “Pauses to funding opportunities” tracking page, last updated in June 2026, confirms that BBSRC’s applicant-led routes were paused while the council moved to an “always open” application model, and that the affected opportunities have since reopened. The table below summarises the position reported across these sources.

    Council / route 2026 development Reported driver
    MRC — applicant-led grants Paused, then phased reopening Transition to new UKRI funding model
    BBSRC — applicant-led grants Paused, reopened by mid-2026 Move to an “always open” application system
    EPSRC — selected routes Paused for review Alignment with the three-bucket structure
    STFC — facilities & subscriptions Cost reductions, some projects shelved Rising international subscription costs

    For research offices, the operational consequence is that a scheme’s status cannot be assumed stable from one funding round to the next. Pre-award teams need to check live status on each council’s page rather than relying on a static internal calendar.

    What Should Research Offices Do to Prepare?

    ARMA — the UK’s professional association for research leadership, management and administration, with member representation from across the higher education and broader research sector — has used its Spring 2026 news round-ups to flag exactly this combination of themes: funding volatility, UKRI’s restructuring, and the practical burden falling on institutional research offices. Read together, the signals point to a small number of concrete readiness actions.

    • Track pause and reopening status directly on funder pages rather than relying on cached internal deadline lists, given how quickly routes have reopened or shifted in 2026.
    • Build strategic-alignment framing into proposal support, since a larger share of new funding is routed through the strategic-priorities and innovation buckets rather than open curiosity-driven calls.
    • Plan for a hybrid systems period as UKRI continues migrating applications from its legacy Joint Electronic Submission (Je-S) system to the newer UKRI Funding Service, with some schemes on each platform simultaneously.
    • Prepare for tighter data and identifier requirements, including consistent use of ORCID iDs, as UKRI’s funding service standardises applicant and organisation data.
    • Model cash-flow scenarios against institutional deficit risk, given the Russell Group’s finding that close to half of English providers are already forecasting a 2025/26 deficit.

    None of this requires guessing at UKRI’s intentions. UKRI has published its own budget-allocation explainer and a live pauses-tracking page, and has committed to a single delivery plan for 2026/27 — the readiness task for research offices is to build monitoring habits around those existing, funder-published sources rather than waiting for a single announcement.

    Answer-First Q&A For Research Administrators

    What Is the Role of a Research Administrator in This Transition?

    A research administrator prepares and submits grant applications, tracks budgets, ensures compliance with funder terms, and manages awards through their lifecycle. During UKRI’s 2026 restructuring, that role expands to include monitoring live pause/reopening status across councils and translating shifting funding buckets into realistic advice for principal investigators.

    What Skills Do Research Administrators Need Now?

    Beyond core pre- and post-award skills, 2026 places a premium on funder-monitoring discipline, strategic framing of proposals against national priorities, and comfort operating two parallel application systems at once. Research offices report that data-governance literacy — particularly around ORCID and standardised identifiers — is becoming a distinct, named competency rather than a background task.

    What Is UKRI’s New Funding Service?

    The UKRI Funding Service is the platform UKRI is migrating applicants to from its legacy Je-S system, intended to standardise and simplify submissions across councils. Through 2026 the two systems run in parallel, so research offices must confirm on a scheme-by-scheme basis which platform a given call uses before advising applicants.

    When Will UKRI Publish Its Next Delivery Plan?

    UKRI’s Chief Executive confirmed to Parliament’s Science, Innovation and Technology Committee in March 2026 that a single delivery plan covering the 2026/27 financial year is in preparation. This is expected to give the sector its first detailed, comparable breakdown of spending under the new three-bucket model.

    What This Means for the Sector Going Forward

    The direction of travel is toward a more strategically aligned, outcome-focused UKRI, not a shrinking one — the headline budget is rising even as its composition changes. That distinction matters for research offices: the volatility is concentrated in which routes are open and on what terms, not in an overall retreat from UK research funding.

    Institutions that treat 2026 as a year of active funder-monitoring, rather than a wait for stability to return, are better placed to advise researchers accurately. Research offices that build the habit of checking UKRI’s published pauses tracker and budget explainer directly, and that read ARMA’s sector round-ups for cross-institutional context, will be positioned to respond as the 2026/27 delivery plan clarifies the practical detail behind the bucket model.

  • UKRI AI Research Labs: What They Mean for Grant Applicants

    UKRI has committed up to £60 million to two new AI research labs, SOFAIR and BOLD, led by UCL and the University of Oxford and funded through the Engineering and Physical Sciences Research Council (EPSRC). For grant applicants, the announcement signals a shift toward fundamental, open-source AI research, cross-institution consortium bidding, and a staged, assessment-gated funding model that is likely to shape future UKRI AI calls across all seven research councils.

    UKRI’s AI research labs are strategic, multi-year research centres — the first major investment released under UKRI’s AI Strategic Framework — built to fund high-risk, high-reward fundamental AI research rather than applied product development. The two labs were announced on 23 June 2026, timed to coincide with what would have been Alan Turing’s 114th birthday.

    What are UKRI’s new AI research labs?

    UKRI’s two new labs are the Science of Fundamental AI Research (SOFAIR) Lab, led by UCL, and the British Open-ended Learning and Discovery (BOLD) Lab, led by the University of Oxford. Both are funded through EPSRC, the UKRI council responsible for engineering and physical sciences, and both work across consortium partners rather than a single institution.

    SOFAIR brings together UCL with the universities of Cambridge, Oxford and Edinburgh to develop next-generation open-source AI architectures that can run on widely available hardware, reducing dependence on a small number of proprietary model providers. It is led by Professor David Barber of UCL. BOLD, led by Associate Professor Jakob Foerster of Oxford, works with UCL and Imperial College London on new learning paradigms — human-centred AI, embodied systems and learning without vast centralised compute.

    Feature SOFAIR Lab BOLD Lab
    Lead institution UCL University of Oxford
    Lead investigator Professor David Barber Associate Professor Jakob Foerster
    Consortium partners Cambridge, Oxford, Edinburgh UCL, Imperial College London
    Core focus Open-source AI on accessible hardware New learning paradigms, embodied AI
    Initial funding released ~£8 million ~£8 million
    Doctoral training ring-fence £2 million (min. 10 students) £2 million (min. 10 students)

    Why did UKRI double the investment to £60 million?

    UKRI’s original plan committed £40 million to a single AI research lab. The final announcement doubled the number of labs to two and lifted total committed funding to up to £60 million — a signal that the review panel judged the applicant pool strong enough to fund parallel, competing research directions rather than consolidate around one.

    Each lab initially receives around £8 million, with the remainder of its allocation released only after an assessment in autumn 2026. This staged, gated funding model — rather than a single upfront grant — is itself a departure from how UKRI has typically structured large capital-style research investments, and it is a detail every institution bidding into future rounds should note.

    What does this mean for grant applicants?

    For institutions and principal investigators watching UKRI’s AI funding pipeline, four practical signals stand out:

    • Consortium bids are favoured over single-institution proposals. Both labs span three to four universities, reflecting UKRI’s preference for pooled expertise over isolated centres of excellence.
    • Fundamental, “blue-sky” research is explicitly back in favour. EPSRC’s framing prioritises foundational questions — new architectures, new learning algorithms — over applied deployment work, reversing some of the recent emphasis on near-market translation.
    • Doctoral and early-career funding is ring-fenced, not discretionary. Each lab must support a minimum of ten doctoral students from a dedicated £2 million allocation, giving PhD applicants and postdocs a concrete, quantifiable funding route rather than a vague promise of “opportunities.”
    • Continuation funding is now conditional on a mid-point review. Applicants should expect future large AI awards to build in an autumn-style assessment gate before releasing full committed funding, rather than disbursing the full sum at award.

    Open-source commitments and spin-out support are also written into both labs’ remits, so institutions preparing future bids should be ready to demonstrate a route to commercialisation and public dissemination of tooling, not only publication output.

    How does this fit UKRI’s cross-council AI strategy?

    EPSRC administers the SOFAIR and BOLD awards, but the labs sit inside a wider, cross-council AI programme. UKRI comprises seven research councils, and the AI Strategic Framework explicitly positions AI as a cross-cutting priority rather than an EPSRC-only remit. The Science and Technology Facilities Council (STFC) — best known for its Hartree Centre, which provides supercomputing and applied AI infrastructure to UK researchers and industry — sits alongside EPSRC as part of that same seven-council structure, underscoring that compute infrastructure and fundamental AI research are being coordinated as complementary strands of one strategy rather than funded in isolation.

    Professor Charlotte Deane, Senior Responsible Owner for the UKRI AI Programme and Executive Chair of EPSRC, described the labs as backing “the bold, high-reward ideas that can shape the future of AI” — language that maps directly onto UKRI’s broader AI strategy documents, which call for regional clusters, sovereign compute capability and closer links between fundamental research councils and applied infrastructure providers such as STFC and the Alan Turing Institute.

    Frequently asked questions

    What is the UKRI AI strategy plan?

    UKRI’s AI strategy sets out a plan to turn the UK’s research strengths into economic advantage by backing fundamental AI research, regional innovation clusters and sovereign AI capability. The SOFAIR and BOLD labs are described as the first major investment released under this framework, with further AI calls expected to follow the same fundamental-research emphasis.

    How much funding did UKRI commit to the new AI labs?

    UKRI committed up to £60 million across the two labs, doubling an original £40 million single-lab plan. Each lab receives an initial tranche of roughly £8 million, with the remaining funds released after an autumn 2026 progress assessment.

    Can researchers outside UCL and Oxford apply for lab funding?

    The two labs are led by UCL and Oxford with named consortium partners, so direct lab leadership is fixed. However, doctoral studentships, postdoctoral posts and collaboration opportunities are expected to open as the labs recruit, and researchers should monitor EPSRC and UKRI funding-opportunity pages for associated calls.

    Which universities lead UKRI’s new AI research labs?

    UCL leads the SOFAIR Lab with Cambridge, Oxford and Edinburgh as partners; the University of Oxford leads the BOLD Lab with UCL and Imperial College London as partners. Both labs also engage with the Alan Turing Institute and UKRI’s existing AI research hubs.

    Implications and outlook

    For research administrators drafting institutional AI strategies, the SOFAIR and BOLD awards are a useful template for what UKRI now expects from a competitive large-scale AI bid: multi-institution consortia, an explicit open-source or public-good deliverable, a ring-fenced doctoral-training component, and acceptance of a staged funding gate rather than a single lump-sum award.

    The autumn 2026 assessment point is worth diarising directly — it will be the first public test of whether UKRI’s gated model releases the remaining funds smoothly or triggers a public renegotiation, and either outcome will inform how future large AI calls from EPSRC, and potentially jointly with STFC’s compute-infrastructure remit, are structured. Institutions preparing bids for the next round of UKRI AI funding should treat this announcement as the current baseline for what “fundable” fundamental AI research now looks like in the UK.

  • UK R&D Investment 2026: Grant Pipeline Impact for Institutions

    The UK government has committed £55.4 billion in detailed Department for Science, Innovation and Technology (DSIT) research and development allocations for the financial years 2026/27 to 2029/30, part of a wider £58.5 billion real-terms R&D budget for the period. UK Research and Innovation (UKRI) alone will deliver £38.6 billion of that total, rising to nearly £10 billion a year by 2029/30. For research administrators, the practical question is not the headline figure but what it does to grant pipelines, funder call volumes, and institutional planning cycles over the next four years.

    UK research and development investment 2026 refers to the multi-year DSIT and UKRI budget settlement, published 30 October 2025, that sets funding envelopes for UK public research bodies from 2026/27 through 2029/30. This is a spending-review allocation, not a single-year grant round — and that distinction shapes how institutions should plan.

    Contents

    What is changing in UK R&D investment for 2026?

    DSIT’s overall R&D budget will grow in real terms across the current Spending Review period, reaching £58.5 billion between 2026/27 and 2029/30, according to the department’s published R&D plans. Of that total, £55.4 billion has been allocated in detail across named organisations and programmes, with the remainder — covering programmes still being finalised — to be confirmed later.

    This is a multi-year settlement rather than a single Budget announcement. It replaces annual uncertainty with a four-year envelope, which changes how institutions can realistically plan grant-writing capacity, co-investment commitments, and infrastructure bids.

    How is the £55.4bn allocation broken down by organisation?

    UKRI is the largest single recipient, with an expected £38.6 billion across the four years — its budget rising from £8.811 billion in 2025/26 to £9.986 billion in 2029/30. A more detailed breakdown of UKRI’s own council-level budgets followed in December 2025. The remaining allocation is split across UK contributions to EU programmes (including Horizon Europe and its successor), the UK Space Agency, the Met Office, the Advanced Research and Invention Agency (ARIA), the National Academies, the Office for Life Sciences, the National Measurement System, and the AI Security Institute (AISI).

    Organisation / programme 2025/26 (£m) 2029/30 (£m) Total 2026/27–2029/30 (£m)
    UK Research and Innovation (UKRI) 8,811 9,986 38,586
    UK contribution to EU programmes 2,736 2,200 8,716
    UK Space Agency 668 720 2,798
    Met Office 310 347 1,467
    ARIA 184 400 1,220
    Office for Life Sciences 129 146 925
    National Academies 217 235 910
    National Measurement System 130 145 558
    AI Security Institute (AISI) 66 60 240

    Source: DSIT, “Research and Development (R&D) plans to 2029/2030”, published 30 October 2025. Figures are planning allocations and, per DSIT’s own disclaimer, subject to in-year reallocation under its new “agile” budget-management approach.

    What does this mean for institutional grant pipelines?

    A rising, multi-year UKRI envelope — from £9.220 billion in 2026/27 to £9.986 billion by 2029/30 — gives research offices a firmer basis for forward-loading grant pipelines than the single-year settlements common in prior spending rounds. Institutions can use the published trajectory to model realistic co-investment and matched-funding exposure three to four years out, rather than reacting to annual uncertainty.

    DSIT itself frames its new approach as more agile: funding can move across financial years where projects are delayed, be reallocated where a programme underspends, or be deprioritised where it is not delivering. That flexibility cuts both ways for pipeline planning — a confirmed four-year envelope is more predictable in total, but individual scheme budgets within it may shift year to year as DSIT and UKRI reprioritise.

    • Build grant-pipeline forecasts around the confirmed multi-year UKRI trajectory, not single-year headline figures.
    • Track the UK contribution to EU programmes carefully — it falls from £2.736 billion (2025/26) to £2.121 billion (2026/27) as Horizon Europe Guarantee costs wind down, which affects institutions relying on guarantee-funded EU collaborations.
    • Monitor ARIA’s rapid growth (from £184 million to £400 million across the period) as a distinct, high-risk/high-reward funding route worth separate pipeline tracking from mainstream UKRI council schemes.

    Will funder call volumes and success rates change?

    UKRI’s own overall research and innovation budget is confirmed as rising during this Spending Review period, reaching almost £10 billion annually by 2030. A rising overall envelope does not automatically translate into proportionally more open calls — much depends on how UKRI’s nine councils allocate the increase between responsive-mode schemes, strategic priority programmes, and infrastructure.

    DSIT has stated three explicit R&D priorities guiding allocation: protecting curiosity-driven, foundational science; supporting strategic government and societal priorities; and targeting innovative, UK-based company scale-up and growth. Research administrators should expect call volumes to grow unevenly across these three streams rather than uniformly across all disciplines.

    Answer-first questions research administrators are asking

    What is the UK government’s total R&D budget for 2026/27?

    DSIT’s overall R&D budget totals £58.5 billion across 2026/27–2029/30, with £55.4 billion of that detailed by organisation in DSIT’s published plans as of 30 October 2025. UKRI is the largest single component, at £38.6 billion over the same four years.

    How much has UKRI’s budget increased?

    UKRI’s budget rises from £8.811 billion in 2025/26 to £9.220 billion in 2026/27, reaching £9.986 billion by 2029/30 — an increase of roughly £1.18 billion, or 13%, across the Spending Review period, per DSIT’s published allocation table.

    Why is the UK’s EU programme contribution falling?

    The UK’s contribution to EU programmes, including Horizon Europe, drops from £2.736 billion in 2025/26 to £2.121 billion in 2026/27 as Horizon Europe Guarantee costs wind down and the UK’s automatic correction mechanism under its association agreement takes effect, per DSIT.

    What return does public R&D investment generate?

    DSIT’s research states that each pound of public R&D investment leverages, on average, £2 of private R&D investment and generates £8 of net benefit in the long run, citing its “Value of Public R&D” research report (DSIT 2025/036).

    What should research offices do now?

    Institutional research offices, grants teams, and pro-vice-chancellors for research should treat this settlement as a four-year planning input rather than a one-off announcement.

    1. Re-baseline internal grant-pipeline forecasts against the confirmed year-by-year UKRI figures rather than the single £55.4 billion or £38.6 billion headline totals.
    2. Flag EU-programme-dependent projects for early review given the falling Horizon Europe Guarantee allocation.
    3. Build ARIA into distinct pipeline tracking, separate from mainstream UKRI responsive-mode schemes, given its faster proportional growth.
    4. Watch for UKRI’s own council-level budget detail and DSIT’s in-year reallocation decisions, since both bodies have flagged an “agile” approach that can shift funding between years and schemes.

    Outlook: the next four years

    The confirmed trajectory to nearly £10 billion in annual UKRI funding by 2029/30, inside a wider £58.5 billion DSIT envelope, gives UK research administration a rare degree of multi-year visibility. The practical work now shifts from interpreting the headline figure to modelling how each council, scheme, and international programme within it will move — a task best served by revisiting institutional grant-pipeline forecasts against DSIT’s published tables rather than summary press coverage. Sound research administration practice in this period means tracking these allocations at the same granularity DSIT itself now publishes them.

  • UKRI Funding Buckets Explained for Grant Holders

    UKRI funding buckets are the four categories — curiosity-driven research, strategic government and societal priorities, supporting innovative companies, and a cross-cutting “enabling and strengthening UK R&D” layer — into which UK Research and Innovation now allocates its £38.6 billion 2026–2030 budget. The model replaces council-by-council settlements with outcome-led pots, and it will shape how every future grant call is designed through to the 2030 spending review deadline.

    UK Research and Innovation (UKRI) is the UK’s largest public funder of research and innovation, distributing money through seven research councils, Research England and Innovate UK. From April 2026, UKRI directs the majority of that money through the new bucket structure rather than through traditional per-council budget lines — the biggest change to its allocation model since UKRI was created in 2018.

    What are UKRI’s funding buckets?

    A UKRI funding bucket is one of the strategic investment categories UKRI now uses to allocate its budget, replacing the previous practice of setting a fixed annual budget for each research council individually. The Department for Science, Innovation and Technology (DSIT) set out the underlying “three R&D priorities” in its 30 October 2025 spending plans; UKRI applied these to its £38.6 billion allocation in its 17 December 2025 budget explainer.

    UKRI’s own framing names three “priority buckets”:

    • Curiosity-driven, foundational research — applicant-led grants and block grants such as Quality-related (QR) funding to English universities.
    • Strategic government and societal priorities — targeted programmes aligned to the government’s Modern Industrial Strategy sectors and wider missions.
    • Supporting innovative companies — commercialisation, knowledge exchange and business scale-up funding, delivered mainly through Innovate UK.

    A fourth, cross-cutting category — enabling and strengthening UK R&D — funds the infrastructure, talent and institutes underpinning all three priority buckets. UKRI does not brand it a fourth “priority,” but it has its own budget line (Table 9 of the explainer), and sector analysts describe it as functioning as a de facto fourth bucket.

    How much money sits in each bucket?

    UKRI’s 17 December 2025 explainer publishes exact four-year totals for the 2026–27 to 2029–30 spending review (SR) period, broken down by bucket:

    Bucket SR-period total What it funds
    1. Curiosity-driven research £14.5 billion Applicant-led grants (£3.3bn), QR funding to universities (£8.9bn), institutes and open-access infrastructure (£2.3bn)
    2. Strategic government and societal priorities £8.3 billion Industrial Strategy sector programmes (£4.5bn), the R&D Missions Accelerator (£500m), the Edinburgh supercomputer (£750m)
    3. Supporting innovative companies £7.4 billion Innovate UK-led commercialisation, HEIF, the Local Innovation Partnership Fund (£440m)
    4. Enabling and strengthening UK R&D £8.4 billion Institutes (£1.6bn), collective talent/doctoral funding (£3.5bn), infrastructure (£2.1bn)

    Adding all four lines gives UKRI’s full four-year settlement of £38.586 billion, rising from £9.22 billion in 2026–27 to £9.99 billion in 2029–30 — the “near-£10 billion” annual run-rate UKRI and sector commentators now reference for the end of the spending review period.

    Three buckets or four? Why the count matters

    The discrepancy between “three priority buckets” and a widely reported “fourth bucket” is not a labelling quibble — it changes how grant holders should read UKRI’s own communications. UKRI’s explainer text still says “investment in three priority buckets,” yet the same document allocates £8.4 billion to a separate, numbered budget line (bucket four) that funds infrastructure, skills and institutes.

    Wonkhe’s analysis of the settlement described this fourth line plainly: it is “what is basically a fourth bucket.” For grant holders, infrastructure and doctoral/talent funding — underpinning every council’s ability to deliver — now sit in a separately governed pot rather than inside familiar discipline-specific budgets.

    What the restructuring means for grant holders

    For applicants and research-office staff, the bucket model changes both what gets funded and how decisions get made:

    • Fewer council-specific figures. UKRI states “a breakdown by research council is only possible for curiosity-driven research” — buckets two and three report only by Industrial Strategy sector, not by funding council.
    • Leverage expectations. UKRI targets “an average leverage ratio of at least £3 of private investment for every £1 of public investment” across strategic and innovative-company calls — expect this built into co-funding criteria.
    • Cross-council, SRO-led programmes. Buckets two and three are delivered through cross-UKRI programmes, each led by one executive-chair Senior Responsible Owner, so calls increasingly span disciplines under one programme brand.

    Curiosity-driven applicant-led research — most familiar to individual investigators — keeps its existing per-council structure and, per UKRI, sees “increases over the period” for every council. Grant holders in Industrial Strategy-adjacent fields (AI, quantum, life sciences, clean energy) should expect more programmatic, mission-shaped calls; those in curiosity-driven disciplines should expect process continuity, a larger overall pot, and some coherence-driven reallocation, such as the planned phase-out of non-recurrent Research England funds from 2027–28.

    Is UKRI cutting STFC and other councils, or just hiding the numbers?

    Search interest in “STFC cuts” reflects genuine sector anxiety, but the honest answer is that the bucket model makes council-level comparisons largely unverifiable from UKRI’s public explainer alone. The only research-council figure UKRI publishes is for applicant-led research within bucket one: the Science and Technology Facilities Council (STFC) gets £344 million of the £3.3 billion four-year applicant-led total, against £1,170 million for the Engineering and Physical Sciences Research Council and £453 million for the Medical Research Council.

    Everything STFC receives through buckets two, three or four — including large-scale infrastructure and international subscriptions, historically a significant share of its budget — is folded into cross-cutting or Industrial Strategy totals with no council attribution. This has drawn direct parliamentary scrutiny: in March 2026, UKRI chief executive Professor Sir Ian Chapman wrote to the Commons Science, Innovation and Technology Committee, giving the first detailed comparison of past and future spending under the new model and describing the shift as “not a simple reclassification” but a “fundamental change in how money flows through the organisation.” Committee chair Dame Chi Onwurah said such comparisons are “especially important for understanding what’s changing and for holding UKRI to account — particularly amid reports of research funding cuts.”

    In short: no UKRI document currently states that STFC’s overall funding is being cut, but no UKRI document currently lets anyone outside UKRI verify the opposite either — which is precisely the accountability gap the parliamentary committee is now pressing UKRI to close.

    Common questions about UKRI’s funding buckets

    How many funding buckets does UKRI actually have?

    UKRI names three priority buckets — curiosity-driven research, strategic government and societal priorities, and supporting innovative companies — plus a fourth budget line, enabling and strengthening UK R&D (£8.4 billion), which sector commentators treat as a de facto fourth bucket.

    Which UKRI funding bucket is the largest?

    Curiosity-driven research is the largest bucket at £14.5 billion over the 2026–2030 spending review period, covering applicant-led grants, Quality-related (QR) university funding, and research institutes and infrastructure supporting fundamental discovery.

    What is UKRI’s leverage ratio target for strategic funding?

    UKRI is targeting an average of £3 of private investment for every £1 of public investment across its strategic government and innovative-company buckets, with higher ratios expected specifically for programmes supporting innovative companies.

    Does the new model change how applicant-led grants are assessed?

    No — applicant-led research keeps its existing research-council structure, with every council seeing budget increases over the period; the bucket changes mainly affect strategic and industrial-strategy-linked programmes, not investigator-led applications.

    Outlook: the road to a near-£10bn annual budget

    UKRI’s annual budget rises steadily across the spending review period — from £9.22 billion in 2026–27 to £9.99 billion in 2029–30 — placing it on a trajectory toward, though not quite reaching, £10 billion a year by decade’s end. UKRI has said it will publish a single delivery plan for 2026–27 in spring 2026, and will continue “increasing the coherence of its portfolio within and across buckets” as the model beds in.

    For grant holders, the practical task now is to map pipeline applications onto the new bucket structure, track which Industrial Strategy programmes intersect with their discipline, and watch the parliamentary committee’s scrutiny sessions for council-level detail UKRI’s own explainer does not yet provide. Institutions with dedicated research administration teams are best placed to translate these bucket-level signals into concrete guidance for principal investigators preparing 2026–27 and 2027–28 calls.

  • Research Professional News Closure: Next Steps

    Research Professional News closure is confirmed: Clarivate will discontinue the publication on 31 December 2026, citing the long-term sustainability challenges of running a specialist journalism business. For UK research offices, that removes a three-decade primary-source channel for research policy and funding news, and the gap needs a deliberate replacement plan built from professional-body, funder-direct and independent-commentary sources — not a single like-for-like substitute.

    Research Professional News is the Clarivate-owned trade publication — successor to Research Fortnight and Research Europe — that has covered UK and European research policy, funding calls and higher-education strategy since the 1990s.

    What Is Happening to Research Professional News?

    Clarivate confirmed in a statement published on its Academia & Government portfolio pages that Research Professional News will discontinue publication on 31 December 2026. The company says it is exploring options with third parties to host the archive, but has committed that past content “will remain available and freely accessible to readers and the wider research community” regardless of whether such an agreement is reached.

    Emmanuel Thiveaud, quoted in Research Information’s coverage of the announcement, described three decades of “authoritative, high-quality reporting and analysis across the complex landscape of research policy, funding and higher education” and thanked the editorial team by name in the closure notice. The final edition publishes on 31 December 2026; UK, European and Funding Insight coverage continues on a business-as-usual basis until then.

    Why Is Clarivate Closing Its Research Journalism Arm?

    Clarivate frames the decision as strategic reallocation, not financial distress in its wider business. The stated rationale is a shift to focus the Academia & Government portfolio on “scalable data, insights, analytics, workflow solutions and expert services” rather than editorial journalism.

    This is a narrower move than it first appears. Clarivate’s funding-opportunities database, Pivot-RP, and its Web of Science analytics products are not part of the closure — only the specialist-journalism arm is affected. Research offices should not read this as Clarivate exiting research intelligence; it is exiting one product line within it, a distinction most coverage of the announcement has not spelled out.

    • Final publication date: 31 December 2026
    • Archive: committed to remaining free and publicly accessible
    • Unaffected: Pivot-RP funding database and Web of Science analytics
    • Affected: original editorial reporting, Funding Insight, and UK/Europe policy news

    Which Sources Fill the Research-Intelligence Gap?

    No single outlet replicates Research Professional News’s combination of daily policy reporting, funder-call tracking and sector commentary. Research offices need a three-tier stack: a professional body for practitioner context, funders’ own direct channels for primary-source accuracy, and independent commentary for cross-sector analysis.

    Source Type Coverage Best used for
    ARMA (Association of Research Managers and Administrators) Professional body UK Practitioner news round-ups, CPD, regional network updates
    UKRI direct channels (council e-alerts, Funding Finder) Funder-direct UK Primary-source funding calls and mandate changes
    Wonkhe Independent commentary UK Daily higher-education policy analysis and politics
    EARMA (European Association of Research Managers and Administrators) Professional body Europe Horizon Europe and pan-European funding context
    INORMS Global network International Cross-jurisdiction benchmarking of research-management practice

    ARMA’s UK membership network already circulates sector news and event round-ups to research managers, making it the closest practitioner-facing analogue to Research Professional News’s community role. UKRI’s own site and council-level e-alerts remain the authoritative first-party source for funding-call detail, since a trade publication was always a secondary layer over that primary data. For the wider policy debate — REF, research culture, institutional strategy — Wonkhe‘s daily briefing already covers much of the ground Research Professional News’s opinion and analysis sections held.

    How Should Research Offices Build a Replacement Workflow?

    Treat this as a migration project with a fixed deadline, not a passive wait for 31 December 2026. Research offices that rely on Research Professional News alerts for grants, mandates or REF-adjacent policy should act on three horizons.

    • Immediate: archive or bookmark any internally cited Research Professional News articles now, since URL structures may change if a third party takes over hosting.
    • Short term: subscribe research-office staff to ARMA’s news channels and set up UKRI council-specific e-alerts and Funding Finder saved searches to replace daily funding-call tracking.
    • Medium term: add Wonkhe’s daily briefing (and, for European-facing offices, EARMA’s newsletter) to cover the analytical and cross-sector commentary that funder-direct channels do not provide.

    Institutions should also audit which internal reports, board papers or policy briefings cite Research Professional News as a source, and flag those for a citation review once the archive’s final hosting arrangement is confirmed.

    Common Questions on the Research Professional News Closure

    Why is Research Professional News closing?

    Clarivate states the closure reflects a strategic refocus of its Academia & Government portfolio onto scalable data, analytics and workflow products, combined with the “long-term sustainability challenges” of running a specialist journalism business inside a data-and-analytics company.

    When does Research Professional News stop publishing?

    The final edition publishes on 31 December 2026. Clarivate has committed that the existing archive will remain freely and publicly accessible after that date, whether or not a third party takes over hosting of the content.

    What should UK research administrators use instead?

    Build a three-source stack: ARMA for practitioner news and community context, UKRI’s direct e-alerts and Funding Finder for primary funding-call data, and Wonkhe for independent cross-sector policy commentary and analysis.

    Is Clarivate exiting research intelligence altogether?

    No. Only the editorial journalism arm is closing. Clarivate’s Pivot-RP funding-opportunities database and its Web of Science analytics products continue unaffected, since the decision targets specialist journalism specifically, not the wider research-intelligence business.

    Outlook for UK Research-Policy Intelligence

    The closure removes a single point of aggregation the UK sector had relied on for nearly three decades, and no successor has consolidated its full scope by the 31 December 2026 deadline. Research offices that build a professional-body-plus-funder-direct-plus-commentary stack now will be better placed than those waiting for a single replacement to emerge, because none is likely to appear.

    For institutions building broader research-administration capability alongside this transition, CASRAI’s research administration resources cover related standards and workflow context.

  • Council of Europe AI Treaty: A Second Track

    The Council of Europe AI treaty — formally the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law — is the first legally binding international treaty on artificial intelligence. Opened for signature on 5 September 2024 and designated CETS No. 225, it runs alongside, not inside, the EU AI Act, and it applies across the Council of Europe’s 46 member states plus a growing list of non-European signatories, including the United States.

    The Framework Convention is a treaty under public international law, not an EU regulation: it creates binding obligations for the states that ratify it, each of which must transpose those obligations into domestic law. That structure makes it a second, parallel governance track for any research institution operating in a Council of Europe member state that sits outside the European Union — the United Kingdom, Switzerland, Norway, Ukraine, Türkiye, and more than a dozen others.

    What is the Council of Europe AI treaty?

    The Framework Convention was negotiated by the Council of Europe’s Committee on Artificial Intelligence (CAI), successor to the ad hoc committee (CAHAI) that began scoping work in 2019. It was drafted by the Council of Europe’s 46 member states with observer states Canada, Japan, Mexico, the Holy See and the United States, plus the European Union and non-member states including Australia, Argentina, Costa Rica, Israel, Peru and Uruguay.

    According to the Council of Europe, 68 representatives from civil society, academia and industry contributed to the drafting. The treaty sets out principles AI systems must respect: human dignity, equality and non-discrimination, privacy and data protection, transparency and oversight, accountability, safe innovation, and remedies for people affected by AI-driven decisions.

    • Deliberately technology-neutral, so the text does not require revision each time a new AI architecture emerges.
    • Applies to AI systems used by public authorities (including private actors acting on their behalf) and by private-sector actors.
    • Excludes national defence matters and most research and development activity, with one important exception (see below).
    • Monitored by a Conference of the Parties, which reviews implementation and facilitates stakeholder hearings.

    Who has signed and ratified it?

    The treaty opened for signature on 5 September 2024. Early signatories included the European Union, the United Kingdom and the United States, alongside several Council of Europe member states. Since then, further states — including Bosnia and Herzegovina, North Macedonia, the Republic of Moldova and San Marino — have signed or moved toward ratification, per the Council of Europe’s treaty tracking page.

    A pivotal step came on 15 May 2026, when the European Union formally ratified the Framework Convention, according to the Council of Europe’s Artificial Intelligence Portal. Ratification does not fold the treaty into EU law; the two instruments remain distinct tracks the Union has committed to enforcing in a complementary way.

    Because the Council of Europe has 46 member states — nearly double the EU’s 27 — a large bloc of countries with binding obligations under this treaty will never be covered by the EU AI Act at all: the UK, Switzerland, Norway, Iceland, Ukraine, Türkiye, Armenia, Georgia, Azerbaijan and the Western Balkan states listed above.

    How does it differ from the EU AI Act?

    The EU AI Act (Regulation (EU) 2024/1689) is directly applicable law inside the EU and EEA, built on a tiered risk-classification system with technical and conformity obligations. The Framework Convention is a different instrument: a human-rights treaty setting baseline principles for ratifying states to implement through domestic legislation, not a self-executing regulatory code.

    Dimension Council of Europe AI treaty EU AI Act
    Legal form International treaty (CETS No. 225) Directly applicable EU regulation
    Territorial reach 46 Council of Europe member states + non-European signatories (US, others) 27 EU member states + EEA
    Approach Principles-based human-rights baseline Risk-tiered technical compliance regime
    Enforcement Domestic implementing law per ratifying state; Conference of the Parties oversight National market-surveillance authorities + EU AI Office
    R&D treatment Excluded, except where testing may interfere with rights/democracy/rule of law Research exemption with narrower conditions

    Legal trackers such as White & Case’s AI Watch note the Framework Convention requires each signatory to ensure remedies are available to those affected by AI systems — an obligation that exists independently of whatever risk category the EU AI Act would assign to the same system.

    Does it apply to research and development?

    This is the detail research institutions most often miss. The Framework Convention does not apply to research and development activities — except when testing of an AI system may interfere with human rights, democracy or the rule of law. The carve-out is narrower than it looks: once a prototype moves from the lab into testing that touches real people’s data or opportunities, the exception can lapse and obligations on transparency, accountability and remedy attach.

    For universities, research funders and multinational consortia, this means AI-enabled research tools — automated peer-review triage, algorithmic grant scoring, participant-recruitment models, predictive analytics on patient or student data — are not automatically outside scope simply because they originate in a research setting. Institutions in non-EU Council of Europe states cannot assume “the EU AI Act doesn’t apply to us” settles the matter; the Framework Convention raises a parallel, sometimes broader, compliance question the moment R&D testing touches real-world rights.

    Frequently asked questions

    What is the Council of Europe AI treaty?

    The Council of Europe AI treaty is the Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, opened for signature on 5 September 2024. It is the first legally binding international treaty requiring signatory states to govern the full lifecycle of AI systems in line with human rights, democratic values and the rule of law.

    Which countries have signed the Council of Europe AI treaty?

    Early signatories included the European Union, United Kingdom and United States, joined since by further Council of Europe member states including Bosnia and Herzegovina, North Macedonia, the Republic of Moldova and San Marino. The EU formally ratified the treaty on 15 May 2026, per the Council of Europe’s official portal.

    Is the Council of Europe AI treaty the same as the EU AI Act?

    No. The EU AI Act is a directly applicable EU regulation limited to the EU/EEA. The Council of Europe treaty is a separate international human-rights instrument spanning 46 member states plus non-European signatories, enforced through each state’s own domestic implementing legislation.

    Does the treaty regulate university and funder AI tools?

    Generally research and development is excluded, but the exclusion lifts once AI testing could affect real people’s rights — for example, algorithmic grant scoring or predictive analytics on participant data. Institutions should not assume research-labelled AI tools sit permanently outside the treaty’s reach.

    Implications and outlook for research institutions

    Research institutions in a Council of Europe member state outside the EU should treat the Framework Convention as an independent compliance track, not a footnote to EU AI Act guidance. Practical steps:

    • Map which AI systems used in research administration, grant assessment or participant-facing services could trigger the treaty’s rights-impact exception once national implementing legislation is adopted.
    • Track ratification status in each jurisdiction where the institution operates, since obligations activate state-by-state, not on a single EU-wide date.
    • Build transparency and remedy mechanisms — notice of AI use, a route to challenge automated decisions — into research-facing AI tools regardless of which regime formally applies yet.

    Globally, the treaty is one entry in a widening, uneven map of AI regulations around the world: the EU’s harmonised regulatory code, the Council of Europe’s rights-based treaty spanning EU and non-EU states, a fragmented patchwork of US state AI laws in the absence of comprehensive federal legislation, and sector-specific rules elsewhere. Institutions tracking several of these regimes at once increasingly need an AI legislation tracker that separates treaty-level from regulation-level instruments, rather than one undifferentiated “AI law” category.

    The Framework Convention will not be the last binding international AI instrument. Its Conference of the Parties is designed to accumulate practice and guidance over time, as other Council of Europe human-rights treaties have. Research institutions that build AI governance around rights-based principles now — not only EU AI Act risk tiers — will be better placed as more states ratify and more domestic implementing laws take effect. CASRAI’s research administration resources track how such cross-border compliance obligations intersect with day-to-day research operations.

  • Tri-Agency Research Data Management Policy 2026

    Canada’s Tri-Agency Research Data Management Policy requires postsecondary institutions and research hospitals that administer funds from CIHR, NSERC and SSHRC to publish an institutional research data management (RDM) strategy, to attach data management plans (DMPs) to specified grant applications, and to prepare for a phased-in data deposit requirement. Launched in March 2021, it is Canada’s first cross-agency RDM mandate.

    The Tri-Agency Research Data Management Policy is a joint funder mandate issued by the Canadian Institutes of Health Research (CIHR), the Natural Sciences and Engineering Research Council of Canada (NSERC) and the Social Sciences and Humanities Research Council of Canada (SSHRC) that sets out institutional and researcher obligations for managing publicly funded research data.

    What is the Tri-Agency Research Data Management Policy?

    The policy was announced jointly by CIHR, NSERC and SSHRC on 18 March 2021, following consultation on a 2018 draft. Its stated objective is to support Canadian research excellence by promoting sound RDM and data stewardship practices across the postsecondary and hospital-based research sectors that receive federal funding.

    Unlike a single-agency requirement, it applies uniformly across all three funding councils, making it the first unified Canadian federal RDM mandate. The agencies chose incremental implementation rather than a single compliance date, phasing obligations in over several years to give institutions and researchers time to build capacity.

    What are the policy’s three pillars?

    The policy rests on three distinct requirements, each with its own timeline and audience. Together they move Canadian-funded research toward the FAIR principles — Findable, Accessible, Interoperable and Reusable — as defined by SSHRC’s guidance for applicants.

    • Institutional strategies: eligible institutions had to develop and publicly post an RDM strategy, then formally notify the agencies of completion by 1 March 2023.
    • Data management plans (DMPs): researchers applying to an initial, agency-specified set of funding opportunities must submit a DMP describing how project data will be collected, documented, stored and shared.
    • Data deposit: grant recipients must eventually deposit digital research data, metadata and code that directly support published conclusions into a recognised repository; the agencies are phasing this in based on the sector’s readiness rather than enforcing it on a fixed date.

    The policy also embeds Indigenous data governance explicitly: institutional strategies must recognise the data sovereignty of First Nations, Métis and Inuit communities, with SSHRC pointing applicants to the First Nations OCAP® principles and the CARE Principles for Indigenous Data Governance.

    How do CIHR, NSERC and SSHRC requirements differ?

    All three agencies operate under one shared policy text, but they do not require DMPs on the same grants. Each agency independently designates which of its own funding opportunities carry a mandatory DMP, published on the shared Science.gc.ca research data management page rather than in a single combined list.

    • CIHR has applied DMP requirements to specific health-research competitions, reflecting added sensitivity around personal health information and research ethics board obligations.
    • NSERC has targeted DMPs at select discovery and strategic programmes in the natural sciences and engineering.
    • SSHRC requires DMPs for designated social sciences and humanities opportunities and publishes the most detailed applicant-facing guidance, including a section-by-section drafting template covering data collection, documentation, storage, sharing, responsibilities and legal compliance.

    Institutions are expected to track which of their researchers’ target competitions are in scope, since the DMP obligation is opportunity-specific rather than blanket across every Tri-Agency grant.

    How does it compare with UKRI, NSF and Horizon Europe?

    Canada’s approach sits between the narrower, opportunity-specific model used historically in the UK and the near-universal mandates now standard in the United States and the European Union. The table below sets out the structural differences institutions moving between these funding systems need to track.

    Framework Steward DMP scope Data deposit approach
    Tri-Agency RDM Policy CIHR / NSERC / SSHRC (Canada) Required only for agency-specified funding opportunities Phased in based on sector readiness; not yet universal
    UKRI Common Principles on Data Policy UK Research and Innovation, across its constituent councils Expected for research councils such as MRC, NERC and EPSRC, per council-specific policy Data expected to be made available and accessible at the point of publication
    NSF Data Management Plan requirement US National Science Foundation A DMP has been mandatory for every NSF proposal, across all directorates, since 2011 Sharing plan required; no single universal deposit mandate across directorates
    Horizon Europe Model Grant Agreement European Commission DMP mandatory for participating projects, typically due by month 6 Open access to research data “as open as possible, as closed as necessary”

    The practical distinction for institutions with international collaborators is scope: NSF and Horizon Europe treat the DMP as a near-default project requirement, while the Tri-Agency policy and UKRI’s council-by-council approach both still gate the DMP requirement to specific competitions rather than every grant.

    What must institutions do to comply?

    Institutional research offices carry most of the compliance burden, since the policy places the strategy obligation on the institution rather than the individual researcher. Compliance work typically covers four areas.

    • Publish and maintain an institutional RDM strategy on a publicly accessible page, with a named contact for enquiries.
    • Build institutional capacity: training, data storage infrastructure, and support for researchers drafting DMPs, often via the DMP Assistant tool operated by the Digital Research Alliance of Canada.
    • Track which specific CIHR, NSERC and SSHRC funding opportunities carry a mandatory DMP so applicants are not caught unprepared at submission.
    • Prepare repository infrastructure and researcher guidance ahead of the phased data deposit requirement, including institutional or national options such as the Federated Research Data Repository.

    Institutions that have not yet published a strategy remain out of step with a requirement the agencies set for 1 March 2023, which is a governance gap research offices should treat as a priority remediation item.

    Frequently asked questions

    When did institutions have to publish their Tri-Agency RDM strategy?

    Institutions eligible to administer CIHR, NSERC or SSHRC funds were required to develop, publicly post and notify the agencies of their institutional RDM strategy by 1 March 2023. This is the only fixed compliance date within the otherwise incrementally phased policy.

    Does every Tri-Agency grant require a data management plan?

    No. Each agency designates its own initial set of funding opportunities that require a DMP at application; the requirement is not blanket across all CIHR, NSERC or SSHRC competitions, so applicants must check the specific programme guidelines before submitting.

    What do institutions need to know about FAIR data under the policy?

    SSHRC’s applicant guidance directs researchers to manage data, where ethically and legally possible, according to the FAIR principles — Findable, Accessible, Interoperable and Reusable — while explicitly noting that grant recipients are not required to openly share data if legal, ethical or Indigenous data sovereignty obligations prevent it.

    How does the policy treat Indigenous research data?

    Institutional strategies must recognise Indigenous data sovereignty, and SSHRC points applicants to the First Nations OCAP® principles and the CARE Principles for Indigenous Data Governance when data involves First Nations, Métis or Inuit communities and their collections.

    Implications and outlook

    For institutions, the Tri-Agency policy converts research data management from a discretionary practice into a governance obligation with a named public strategy, a training mandate and eventual deposit infrastructure requirements. Research offices that treat the 2023 strategy deadline as complete, rather than as a living document, risk falling behind as the agencies phase in data deposit.

    For researchers collaborating internationally, the comparison with UKRI, NSF and Horizon Europe matters operationally: a DMP built for an NSF-funded partner project, where a plan is mandatory for every proposal, will not automatically satisfy a Tri-Agency opportunity where the DMP requirement is competition-specific — and vice versa. Institutions running multi-funder projects should map DMP and deposit obligations per funder rather than assuming one plan transfers across systems.

    As Canada’s data deposit requirement moves from phased design toward implementation, institutions with mature repository infrastructure and clear researcher guidance will be better positioned than those still relying solely on their 2023 strategy document. This sits within the wider discipline of research administration, where funder RDM mandates increasingly intersect with data governance, ethics review and open-access policy.