Artificial Intelligence Humanities Sandpit Grants 2026 (Canada, UK and US):

Artificial Intelligence Humanities Sandpit Grants 2026 (Canada, UK and US): Closed Invite-Only Research Funding

The Artificial intelligence humanities sandpits: Canada, UK and US invite only opportunity was a closed research grant call run through UK Research and Innovation. The lead funder was the Arts and Humanities Research Council (AHRC), with the Engineering and Physical Sciences Research Council (EPSRC) as a funder and the Social Sciences and Humanities Research Council of Canada (SSHRC) as co-funder.

This page is now a historical reference. The official UKRI Funding Finder page marks the opportunity Closed. AHRC had to receive applications by 28 May 2026 at 4:00pm UK time, and the page says applicants could not apply after that time. The official page does not announce a subsequent round, a replacement deadline, or a new invitation process. There is therefore no current application date to add here.

What the 2026 call funded

The call supported research that placed humanities insights and methodologies at the heart of artificial intelligence technology design. It was not a general-purpose AI grant and it was not intended for a technical project with a short ethics section added at the end. UKRI required meaningful collaboration between humanities or arts scholars and technical AI researchers.

The funder wanted projects that could make a concrete contribution to interpretive AI. Examples on the official page included functioning prototypes with measurable improvements in interpretive capabilities; evaluation frameworks and benchmarks for cultural sensitivity, contextual awareness, or perspectival reasoning; training or fine-tuning approaches that represent multiple perspectives and cultural contexts; modular representation methods; and technical components that operationalise humanities concepts or methods.

The emphasis was mutual influence. A strong project would show how humanities or arts methods changed technical design, while technical work also tested, extended, or challenged interpretive frameworks. Relevant work could involve hermeneutic, aesthetic, or narrative reasoning; ambiguity handling; plurality-preserving outputs; or computational approaches that turn a humanities concept into an implementable model component, evaluation method, or design practice.

The official guidance also asked applicants to connect the proposed work to the sandpit’s key challenges, explain why the activity was feasible at the proposed scale, and identify potential real-world use cases. Applications needed to make planned outputs clear, explain how those outputs would produce outcomes, and describe the impacts they might generate. The programme did not promise to fundamentally redesign AI systems in one project or group of projects; its stated purpose was to lay groundwork for longer-term innovation.

Funding available

UKRI listed a £1,160,000 total fund and a maximum award of £262,250. It expected to fund up to five grants. The headline national allocations were:

UK-based team members: £780,000 was available across the funded teams. The full economic cost of an individual project could be up to £156,250, with AHRC funding 80% of that FEC, or £125,000.
Canada-based team members: CAD$1,000,000 was available for Canada-based teams. The detailed funding section set a maximum of CAD$200,000 per project.
US-based team members: US researchers could be funded from the UK or Canada budgets, subject to the opportunity’s rules and the structure of the project team.

The UK and Canadian figures should not be added together as though they were a single unrestricted pot available to every applicant. They describe national funding arrangements, and the final request had to follow the relevant funder’s cost rules. The official guidance also stated that Canadian team-member staff costs were not included in the grant amount because salary costs were covered elsewhere, and that SSHRC grant funds must not be used to compensate applicant team members.

The award duration was 18 months. UKRI’s overview said projects had to begin before 1 October 2026 and end before 30 April 2028. The detailed duration section gave the tighter operational wording: projects had to commence on or before 1 October 2026 and be completed by March 2028 at the latest. Anyone using this archive to understand an awarded project should treat the detailed duration wording and the formal award documentation as controlling.

Eligibility for the closed cycle

This was an invitation-only stage after the Montreal sandpit. An applicant could apply only if UKRI or its partners had invited them to do so following attendance at the February workshop in Montreal. The call was not open to a new applicant who had not taken part in that cohort, and it was not a normal open competition where an unrelated team could submit a proposal from scratch.

Applicants had to be based at a UK, Canadian, or US research organisation eligible for funding. The official eligibility wording referred to organisations eligible for AHRC or SSHRC funding. A Canadian international project co-lead had to be affiliated with a Canadian postsecondary institution holding SSHRC institutional eligibility and had to comply with SSHRC’s grant-application regulations, the Tri-agency Guide on Financial Administration, and the policy on sensitive technology research and affiliations of concern.

The UK submission structure mattered. AHRC was a lead funder, and international researchers could be included as project co-leads international, abbreviated PcL (I). Canada-based and US-based applicants were to use that international co-lead role when co-leading alongside UK researchers. Only UK-based applicants could be listed as the project lead in the Funding Service, and only the lead UK research organisation could submit the application to UKRI.

Project teams were expected to consist primarily of sandpit participants. The guidance allowed additional people who had not attended, such as PhD students, research assistants, technicians, or industry partners, but their involvement had to be justified. The proposal needed to explain what specific skills those additional members supplied and why those capabilities were not available within the sandpit cohort.

The research itself also had to be genuinely interdisciplinary. A humanities scholar attached to a conventional AI project would not satisfy the spirit of the call if their contribution had no effect on the design, evaluation, or technical direction. Conversely, a technical group could not treat humanities concepts as decorative language. The collaboration needed a credible route from interpretive or humanistic work to a prototype, benchmark, model component, technical pipeline, or other implementable output.

How applications were submitted

The 2026 grant application used the new UKRI Funding Service, not the Joint Electronic Submissions system. The organisation needed to be registered before the application could move through the service. Because the call was by invitation only, an invited applicant had to use the invitation link to begin.

The official sequence was:

Select Start application on the Funding Finder page.
Confirm that the person completing the form was the project lead.
Sign in or create a Funding Service account. Account creation required selecting the organisation, verifying the email address, and setting a password.
Answer the questions in the text boxes. Applicants could save their work and return later, or work offline and paste the answers back into the service. Where a document upload was requested, the Funding Service supplied the upload instructions.
Check the application in read-only view before sending it to the research office.
Send the completed application to the research office for checking. The office could return it for edits.
The research office submitted the checked application to UKRI. The lead UK research organisation, rather than an individual collaborator, was responsible for the formal submission.

The application had to reach AHRC by 28 May 2026 at 4:00pm UK time. UKRI advised applicants to account for their institution’s internal deadline. After submission, the application could not be changed and submitted applications would not be amended. The official page also warned that an application that did not follow the guidance could be rejected.

Application content and supporting material

The Funding Service questions required more than a short project pitch. The summary had a 550-word limit and asked for the context, challenge, aims and objectives, potential applications, and benefits of the proposed research. The core team section required roles, with only one person listed as project lead. The available role types included project co-lead UK, project co-lead international, specialist, grant manager, professional enabling staff, doctoral student, research and innovation associate, technician, visiting researcher, and researcher co-lead.

The vision section had a 750-word limit. It asked what the team hoped to achieve, which real-world scenarios could benefit, and how the work would improve on existing AI tools or services. Assessors looked for excellent and distinctive work connected to interpretive technologies, genuine mutual enrichment between technical and humanistic methods, clear potential benefits, and a well-supported case for timeliness.

The approach section had a 2,750-word limit. It needed to explain delivery, the added value of international collaboration, the work’s feasibility within the available time and funding, the methodology, delivery risks and mitigations, the research environment, and how outputs would translate into outcomes and impacts. The team also had to address resources and cost justification, ethics and responsible research and innovation, data management and sharing, project partners, trusted research and innovation, and international collaboration.

Canadian participation required the SSHRC budget form supplied with the opportunity. Project partners had to be named with their organisation details, contact details, the type and monetary value of their contribution, and a description of their role. Where partners were included, the application required a single PDF containing their letters or emails of support. Each letter or email had to confirm commitment, explain the value and relevance of the work, describe the partner’s additional contribution, and stay within the page limit set by UKRI.

What reviewers were looking for

The assessment areas were vision, approach, applicant and team capability to deliver, ethics and responsible research and innovation, resources and cost justification, project partners, data management and sharing, and trusted research and innovation. The panel was to be constituted by the Sandpit Director and Mentors and chaired by an independent academic who had not been present at the sandpit. UKRI said it aimed to complete assessment within two months of receiving an application and would provide feedback with the outcome.

The practical lesson from the criteria is that the funded project needed a traceable chain from problem to method to technical change to evidence. Broad claims about making AI more human were not enough. A credible proposal would define the interpretive problem, explain the humanities contribution, identify the technical intervention, set out how the result would be tested, and show why the UK-Canada-US collaboration added something material. It would also be honest about risk, data governance, partner commitments, and the limits of what an 18-month award could establish.

Archive status and official source

This opportunity should not be presented as open. Its official status is Closed, its deadline was 2026-05-28, and the official page does not announce a next cycle. The historicalReference flag in this page’s front matter preserves the real closed-cycle date while allowing the entry to function as an archive record rather than an expired live listing.

For the original funding rules, the authoritative source is the UKRI Funding Finder page:

https://www.ukri.org/opportunity/artificial-intelligence-humanities-sandpits-canada-uk-and-us-invite-only/

The page is administered by UKRI and identifies AHRC and EPSRC as funders, with SSHRC as co-funder. Researchers looking for a future opportunity should verify any new announcement directly with UKRI, AHRC, or SSHRC rather than treating this closed 2026 call as evidence that a new round is available.