Funding opportunities

Dataset creation & challenge projects (Closed)

This call closed to applications on 24 February 2026

This page is for reference purposes only

Collaborations led by UK academic researchers can bid for up to £100,000 for dataset creation projects through this call, which aims to encourage development of new datasets supporting scientific advancement and technological innovation. On this page you can find all details and documentation associated with the call, which closes at 5pm GMT, 24 February 2026.

Please fully review the guidelines and FAQs below before contacting the hub team with questions.

Overview

The availability of rich, diverse, and well-curated datasets is vital for advancing scientific and technological innovation. Dataset challenges serve as a mechanism for benchmarking algorithms, identifying novel solutions, and fostering collaborative research.  

This funding call seeks to support projects that will:

  • Create new, high-quality datasets: identifying gaps in existing data resources, enabling new research avenues and potential for significant impact.

  • Develop multimodal benchmark datasets and tasks with a clear real-world application.

  • Curate and enhance existing datasets: Significantly improving the quality and use of existing datasets.

  • Design and host dataset challenges: Leveraging datasets to host impactful challenges that encourage problem-solving within the research community.

Funding details

The total project cost must not exceed £125,000.

Please note that this grant provides 80% funding (up to a maximum of £100,000). The remaining 20% must be covered by the host institution.

  • Maximum project value (100% eligible costs): £125,000

  • Maximum grant amount (80% eligible costs): £100,000

We will fund:

  • Academic research staff costs that are costed as directly incurred staff.

  • Directly allocated (DA) costs* such as estates, indirect, and infrastructure technician (not researchers’ DA salary contributions) costs that are essential to project delivery

  • Fully justified consumables

  • Access to national and other open access facilities

  • Software licenses

  • Costs associated with research data management and sharing

  • Impact acceleration and knowledge exchange activity

  • Equipment under £25,000 inclusive of VAT

  • Travel and subsistence

  • Overhead costs will be considered on a case‑by‑case basis

    * In line with UKRI's funding principles, costs requested should normally be directly incurred. However, a limited amount of directly allocated costs may be considered for this funding opportunity, if essential to project delivery.

We will not fund:

  • Contributions to researcher’s salary (i.e. directly allocated staff costs)

  • Contract extensions on existing projects

  • Equipment costing more than £25,000 (inc. VAT)

  • Fees or stipends associated with Masters or PhD studentships

  • Open access publishing costs

  • Basic computer equipment for staff already employed by the organisation

  • Bids solely requesting equipment, even if the equipment costs less than £25,000

  • Bids solely requesting travel and subsistence. Other funding opportunities that focus on exchanges and networks may be available from the hub in 2026.

Who can apply

This call is open to academic researchers based at UK institutions eligible for UKRI funding. Following the EPSRC individual eligibility guidelines, fixed-term employees such as PDRAs are eligible as a PI if they can get support from their organisation (e.g. line manager approval) and the term of employment extends beyond the duration of the proposed research project. 

Academics can apply as an individual or collaborative team, from a single or across multiple organisations. We encourage applicants from interdisciplinary teams spanning research areas linked to the hub’s working groups. For teams with multiple organisations, there must be a single lead applicant and institution identified which will submit the application on behalf of all partners. 

We particularly encourage applications from early career researchers, including those at postdoctoral level (with PI approval). We also encourage applications that aim to develop new and existing collaborations between academic and industry partners. 

We are committed to achieving equality of opportunity for all applicants and encourage applications from a diverse range of researchers and will seek to support flexible ways of working as requested.

What we are looking for

While we remain open to a broad range of applications, we are particularly interested in dataset curation, enhancement, or challenges deemed relevant to at least one of the hub’s thematic working groups.

We are also keen to bridge the gap between academia and industry and encourage applications that involve collaboration.

It is essential that your application:

  • Aligns with the mission of the Hub

  • Ensures output quality and realism, and / or mitigates risks and bias

  • Clearly demonstrates real-world impact and improvements to generative AI tools

Please see our funding guidelines for full details.

Application process

**The closing date for this call has passed and we are no longer accepting applications**

Applications should be completed using our Dataset fund application form which is available to download here. The completed form should be sent to the Gen AI Hub by 17:00 (GMT), Tuesday 24 February 2026.

Assessing your application

Applications will be reviewed anonymously by our review panel, with the Hub operations team removing identifiable details before submitting to the panel. If your submission does not follow the guidance provided, it may be rejected.

You can find out more information about the timeline for this call and the assessment process in the funding guidelines and the FAQs.

Key details

  • Total funds available for all projects: £400,000

  • The total costs of your project can be up to £125,000. The Gen AI Hub will fund 80% of eligible costs - up to £100,000.

  • Projects can be up to 12 months in duration.

  • Deadline for applications: 24 February 2026

Documents

**The closing date for this call has passed and we are no longer accepting applications - these documents are for reference only **

Useful Links

FAQs

  • Our purpose is to help to make Generative AI models more customisable, reliable and trustworthy. There are huge benefits for society, science and the economy, if we can make these models easier to interact with and train.

    Our overarching goal is to create tools that UK industry, science and government can use to fine-tune generative models (GMs). We will do this by:

    • Making these tools widely available,

    • Enabling easier interaction with GMs

    Other objectives include:

    1. Facilitating the formation of start-up companies and contributions to open-source projects using GMs.

    2. Enabling academic researchers from a wider range of universities and backgrounds to collaborate with companies, large and small, on impactful research projects.

    3. Developing a better understanding of AI’s impact on society, along with solutions to address common challenges for society. These could include:

    • Creating open-source software to both train large GMs and fine-tune GMs on specialised data;

    • Making GMs more reliable and trustworthy by incorporating constraints and understanding the capabilities of language models’ reasoning abilities;

    • Using GMs to preserve privacy and ensure fairness;

    • Understanding the legal rights regarding user data and responsibilities relating to malfunctioning generations from GMs;

  • The full economic cost (FEC) of your project can be up to £125,000. The Gen AI Hub will fund 80% of the FEC per project, which equates to £100,000 per project.  The remaining 20% should be covered by the applicant institution or through additional funding sources. 

  • There is no requirement for letters of support. However, you do need institutional support and will be asked in the form whether you can provide evidence that your university is supportive of the application.

  • No. Investigators named on the grant may support an application from someone else but may not submit one directly. Members of working groups from the AI Hub in Generative Models can apply if they are not one of the investigators named on the grant.

  • Yes. Investigators from other hubs can apply for this funding.

  • If a PhD student has completed their studies and is no longer in receipt of a stipend or other form of studentship funding, then they are eligible to be employed as a research assistant or project support on an awarded grant. They must have completed their studies.

  • Projects should be up to 12 months in duration and start before January 2027.

  • Travel costs are allowed but bids solely requesting travel and subsistence are not permitted. We will provision a fund for this in 2026, so please look out for this.

  • Equipment costs of up to £25000 (inc. VAT) are allowed, however, only 80% of full costs are recoverable under the rules of this call. Bids solely requesting equipment, even for less than £25,000, are not allowed. Basic computer equipment for staff already employed by the applicant organisation is not permitted.

  • All applications will be assessed on how well they meet the criteria stated in the funding guidelines by an independent review panel, which will include external reviewers and exclude anyone with a declared conflict.

    All applicants will be required to declare relevant conflicts, and we will publish information about awards made, outlining the number of applications and the criteria applications were reviewed against.

  • Requests for overheads will be considered on a case-by-case basis, with priority given to costs that directly support delivery of the work.

  • Projects should be costed at 100% of eligible costs (not exceeding £125,000). The recoverable cost (80%) should not exceed £100,000.

    Estates and Indirect Costs

    To assist institutions with estates, indirect costs and overheads associated with this type of project, the hub has made these eligible where the project relies on it - despite the exclusion of DA staff salaries. These should be calculated using your institution’s standard TRAC methodology.

  • Like many UKRI-funded investments, the hub has specific rules on which of those costs the grant will actually reimburse.

    Treatment of Staff Costs (DI vs. DA)

    The distinction between allowable and non-allowable staff costs is based on the Directly Incurred (DI) and Directly Allocated (DA) classifications:

    Directly Incurred (DI) Staff – ELIGIBLE: the hub explicitly funds "Academic research staff costs." This includes researchers, PDRAs, or technicians dedicated to the project. For Early Career Researchers on fixed-term contracts who are acting as PIs but performing the hands-on research, these costs should be included as DI to be eligible for funding.

    Directly Allocated (DA) Staff – INELIGIBLE: Contributions to the salary of investigators (for example senior oversight/supervisory time) are not funded by the Hub.

  • Post doc researchers should be costed under the directly incurred category. Directly allocated salary costs are not allowed under this funding call. Please speak to your research finance team for help with this.

  • The awarding body would be UCL, although the funding comes to the hub from EPSRC.

What would a model costing look like?

Please see our suggested costing below.

Category Costing Total costs 100% (£) Cost to funder 80% (£)
Directly Incurred - Staff Staff (DI) eg PDRA 29,506.48 23,605.18
Directly Incurred - Non Staff Travel and Subsistence 7,120.00 5,696.00
Directly Incurred - Non Staff Other Directly Incurred eg software & licences, compute, catering, workshops costs 32,800.00 26,240.00
Directly Allocated Staff (DA) eg PI or Co-I - -
Directly Allocated Estate Costs - standard costing by institution 9,787.80 7,830.24
Directly Allocated Infrastructure Technician Costs 1,341.00 1,072.80
Indirect Costs Indirect Costs - standard costing by institution 42,429.00 33,943.20
Total 122,984.28 98,387.42

Any other questions?

We hope the FAQs have helped to answer your questions. If not, please complete this form and we will be in touch shortly.