AI-deas: Do you have an idea for game-changing AI innovations that could deliver public benefit?

ai@cam is the University of Cambridge’s new flagship mission to create AI technologies that serve science, citizens, and society. We’ll be launching a website in September with further information about our work, but in the meantime you can read the review that led to our creation here: https://www.cam.ac.uk/research/aicam Our plan is to develop a programme of activities that help address the issues set out in this review.

Amidst fevered conversations about progress in AI, the foundational work to connect technology development to public value is more important than ever. We face a cluster of complex challenges: recovering economic growth post-Covid; ensuring health and social care systems deliver for vulnerable people and patients; building resilience to the impacts of climate change; tackling pollution and biodiversity loss to enable environmental recovery; improving public service delivery in the context of strained resources; and more. Despite AI’s potential as a tool to help address these challenges, AI R&D is too often disconnected from public benefit.

The University has an ambitious vision for AI that can deliver public benefit, which you can read in our review. To deliver on this vision, we need a new infrastructure for interdisciplinary AI innovation.

Challenge-led research initiatives can galvanise the research community in support of priority issues and help build capability in areas of need. Inspired by these approaches, ai@cam is launching a cross-institutional research initiative focused on how Cambridge can leverage AI to help tackle today’s major scientific and societal challenges. We want to generate new thinking about the role of AI in addressing major challenges, build coalitions in areas of need, and create partnerships to take research from lab to real-world benefit. Research teams supported by the scheme will have access to funding and research support to scale their work and impact. This challenge-led research funding scheme is our first initiative, and we’ll launch further activities during the autumn and winter.

This call is open to staff at the University of Cambridge, working in any discipline, including natural, physical, biological, social, medical, and computer sciences, arts, humanities, and engineering. PhD students can also submit ideas, but please get permission from your supervisor first.

At this stage, we’re looking for concept papers that set out an idea for challenge-led research that could deliver positive real-world outcomes through AI. You don’t need a fully developed research proposal to apply. For example, you might have the idea, but not yet have access to the tools or collaborators to develop and implement it. Alternatively, if you’d like to be involved in developing a challenge but don’t have a clear area of focus, check out the FAQ below about how you can get involved.

Today, we’re launching the first of a two-step process. We’re inviting concept papers with ideas for research that can help deliver ai@cam’s mission. The second step of this process will build research teams in thematic areas, scoping ambitious challenge-led programmes on those areas. The reason for working in two phases like this is to encourage collaboration: we know there is interest in AI from across the University, and world-leading expertise across departments, but it isn’t always easy to find collaborators, and we hope to facilitate this process.

While this first stage of expressions of interest is focused on the research community within the University, we hope the resulting challenges will engage with external partners both in the UK and beyond. We’re particularly interested in engaging with communities affected by the development of AI and user groups with an interest in the design of AI systems. If you’re an external collaborator with an interest in ai@cam, you can sign up for updates about the mission here or reach out to Jess Montgomery (jkm40@cam.ac.uk). Please bear with us while we’re in set-up phase; responses may take a little longer than usual.

After our roadmapping sessions, ai@cam will work with research teams to develop their idea into a challenge-led research initiative. We’ll start by providing start-up funding of up to £150k to help get the initiative off the ground. We expect this funding can be used flexibly, for example for research activities, equipment, staffing, teaching, or network building. We’ll host networking and engagement activities. We’ll provide support developing external connections and partnerships. We’ll also work with you on communications and profile-raising of the work. The exact package of support will be developed in collaboration with research teams as we develop the missions, and will flex over the two year period where we hope to help you launch your initiative.

At this stage, researchers are submitting ideas via the concept papers, with the expectation of a time commitment for workshops later in the project development process. Concept papers will be reviewed by the ai@cam team and clustered into themes. We’ll then bring together researchers within a theme to scope a challenge-led programme. Depending on the volume of interest in this call, this would be through either one or several workshops in the autumn and winter.

We anticipate offering start-up funds of £150k to selected challenges, working with those challenges to generate further income. Depending on the scale of demand, we may also launch a small-projects start-up scheme later this year. We’ll be using the concept paper and challenge scoping process to assess whether there is demand for this.

ai@cam’s mission is to develop AI technologies that serve science, citizens, and society, and our focus is interdisciplinary research that addresses real-world challenges. We’re interested in projects that have a focus on public benefit, are trying to work across the pipeline from research to innovation and implementation, and that bring together expertise from across disciplines and user groups. Concept papers could relate to a new idea or build on an existing collaboration.

Funding will be deployable flexibly depending on the needs of the project. We will ask funding recipients to work with members of the ai@cam team to develop their projects, with a tailored support package. We anticipate working with challenges developed in this round for a two-year period, and for start-up funds to be deployed over this period. There may also be times over this period where we ask funding recipients to contribute to ai@cam events and comms activities.

To submit a concept note, visit https://forms.gle/e6jBHyUdbZ3wBV2f7

We’re looking for a short pitch that sets out your idea for using AI in the service of science, citizens, and society, alongside some details about you.

The application form asks for information about:

• Your project idea (2000 characters).
• Why this area of research is important (1000 characters).
• How you’d approach tackling this challenge (1000 characters).
• Desired outcomes from this work (1000 characters).
• Why Cambridge is positioned to act in this area (1000 characters).
• What type of support you would need to develop this challenge (1000 characters).

If you’re not at the stage where you’d like to submit a concept paper, but you’d like to be engaged in the development of ai@cam’s AI challenges, please get in touch with Jess Montgomery (jkm40@cam.ac.uk) with a brief overview of your interests.

You can contact Jess Montgomery at jkm40@cam.ac.uk. While we’re building the ai@cam team, responses may take a little longer than usual. Neil Lawrence will host a Zoom drop-in session for any questions about the scheme on 19 September from 13:00 – 14:00. To receive Zoom details for this session, please register at: https://cl-cam-ac-uk.zoom.us/meeting/register/tJEkf--upzsvE9PLQUPgSZCGjBlO_dl-0szF