Department for Science, Innovation and Technology: Systemic AI Safety Fast Grants
Overview
Addressing AI's risks to people and society requires looking beyond AI models' capabilities. AI risk management must also include understanding the impact on the systems and infrastructure in which AI models operate.
Systemic AI safety means safeguarding the societal systems and critical infrastructure into which AI is being deployed––to make our world more resilient to AI-related hazards and to enable its benefits.
The AI Safety Institute (AISI), in partnership with UK Research and Innovation (UKRI), is excited to support impactful research that takes a systemic approach to AI safety.
They will begin by offering a round of seed grants, followed by future rounds with more substantial awards.
They expect to provide successful applicants with ongoing support, computing resources, and access to a community of AI and sector-specific domain experts.
They designed their grant process to be maximally inclusive––applying for a grant should be fast and easy.
What they are funding
In partnership with UKRI, the AISI will shortly invite researchers to submit grant proposals that directly address systemic AI safety problems or improve our understanding of systemic AI safety.
They will prioritise applications that offer concrete, actionable approaches to significant systemic risks from AI.
They expect to offer around 20 exploratory or proof-of-concept grants and will invite future bids for more substantial proposals to develop research programmes further.
Eligibility
Because AI presents novel risks to all areas of society, they welcome proposals from researchers in any field. The only eligibility requirements are that you are:
- Associated with a host organisation that can receive the awarded funds. This could be a university, a business, a civil society organisation, or part of government.
- Based in the UK (but teams may involve international collaborators).
What they are looking for in proposals
They want successful proposals to translate into meaningful impact. They aim to build a community of researchers focused on systemic AI safety. To this end, they will meet regularly with grantees and connect them to other stakeholders within and beyond the UK government to ensure that ideas reach their full potential. Specifically:
- They want grantees to meet with them every ~3 months (at a half-day online/hybrid workshop) to discuss their research progress and network with other awardees. These meetings aim to exchange ideas across the different projects and start building a community around systemic AI safety.
- They will work with each grantee team to find partners to implement or test their ideas where appropriate.
Grantees with the most promising projects may be eligible to develop an application for a continuation grant. They will work closely with awardees to identify projects most likely to lead to impactful outputs and help them develop proposals for future funding. We will announce more information about future rounds in due course.
AISI will collaborate on this work with UKRI, The Alan Turing Institute and other AI Safety Institutes worldwide for this programme.