Artificial Intelligence is a rapidly growing and globally transformative technology, with the potential for wide-ranging impacts across all sectors of society and the economy.
Realising positive outcomes from the development, deployment and use of AI systems – outcomes that benefit the economy and society, and support and improve people’s lives – will require input from the widest range of expertise and perspectives.
It will also rely on building public trust, which only a robust responsible AI ecosystem can foster.
Written by BRAID programme directors Professors Ewa Luger and Shannon Vallor, this UK Research and Innovation (UKRI) blog post discussing the importance of the arts and humanities to responsible AI sets the context for the fellowships scheme.
The programme directors aim to leverage the UK’s research strengths in responsible AI to promote innovation, economic growth, and public welfare, and to provide thought leadership for the Arts and Humanities Research Council (AHRC) and UKRI in this space.
Funded by the AHRC, the programme is informed by the UKRI AI Review, Transforming our world with AI (PDF, 5.5MB). It also builds on the AHRC’s strategic priorities and objectives outlined in the National AI Strategy, which highlights ‘delivering responsible AI’ as necessarily central to UKRI’s approach, and notes the interdisciplinary collaboration required to achieve it.
By fostering collaboration between academia and a diverse range of stakeholders that are developing, deploying, or using AI technologies, the programme seeks to inform regulations, standards, and industry practices.
With collaborative research at its heart, we aim to build connections between disciplinary and sectoral actors to incentivise responsible and ethical innovation in the field of AI.
Addressing the challenges involved in the development, deployment and use of responsible AI across myriad societal and sector contexts requires deeply interdisciplinary and collaborative approaches to research. The Bridging Responsible AI Divides (BRAID) Fellowships and our funders at the Arts and Humanities Research Council are determined to embrace this approach.
The field of AI comprises a complex ecosystem of actors, stakeholders, researchers and publics. Yet as explained in our earlier blog post, too often the evaluation of excellence and breakthroughs in the technology sector prioritise scientific and technical innovation over social and ethical innovation.
The first round of fellowships presents a unique opportunity to explore and understand empirically the contributions that the arts and humanities can make to the AI ecosystem and the implications for this kind of research across different applied contexts.
Research methods from the arts and humanities originate in the fields of arts, literature, philosophy, history, and cultural studies yet are invaluable tools for researchers across many disciplinary divides. They emphasise qualitative analysis, interpretation, and subjective understanding, and allow researchers to explore complex human experiences, cultural phenomena, and social contexts. They inform projects like the Advanced Care Research Centre, AI Forensics, Transkribus, Senseful AI and Living with Machines. They are crucial to any technology-enabled world that seeks to support and benefit human flourishing. Yet their utility beyond core subject areas in fields like responsible AI is often underappreciated, underdeveloped and underutilised. It is a particular divide that the BRAID Fellowships promise to bridge.
An overview of common research methods and practices within the arts and humanities makes immediately apparent the promise that the arts and humanities can make to our technologically-enabled world:
- Qualitative Research: Qualitative research methods are commonly employed in the arts and humanities to study subjective experiences, cultural practices, and social phenomena. Techniques such as ethnography, interviews, case studies, and content analysis help researchers gain a deeper understanding of the complexities and nuances within human experiences and cultural contexts.
- Historical Analysis: Historical research involves the critical examination of past events, cultures, and societies. Historians often use primary sources, archival materials, and various historical texts to analyse and interpret historical narratives and reconstruct past events. This method enables researchers to understand the evolution of cultures, ideas, and societies over time.
- Literary Analysis: Literary research involves the critical analysis and interpretation of literary texts such as novels, poems, and plays. Researchers often apply various approaches, including close reading, deconstruction, and cultural criticism, to explore themes, symbols, and narrative techniques within the text. Literary analysis allows researchers to uncover deeper meanings and understand the cultural and social contexts in which the texts were produced.
- Arts-Based Research: Arts-based research methods integrate artistic practices such as visual arts, music, dance, and performance into the research process. Researchers use creative expression and artistic mediums to investigate complex social and cultural issues, often emphasising the use of visual imagery, storytelling, and creative representation to convey research findings and explore diverse perspectives.
- Philosophical Inquiry: Philosophical research involves critical thinking and reasoning to address fundamental questions about existence, knowledge, ethics, and the nature of reality. Philosophers often employ conceptual analysis, argumentation, and logical reasoning to examine complex theoretical and ethical issues, contributing to broader discussions within the humanities.
- Cultural Studies: Cultural studies research focuses on the analysis of cultural practices, identities, and social dynamics within specific cultural contexts. Researchers draw from interdisciplinary perspectives, incorporating methods from sociology, anthropology, and literary studies to investigate the production, circulation, and consumption of cultural artifacts and practices, including media, art, and popular culture.
- Interdisciplinary Approaches: Many research projects in the arts and humanities adopt interdisciplinary approaches that integrate methodologies from multiple disciplines. Researchers combine insights from various fields to explore complex issues, fostering a comprehensive understanding of cultural, social, and historical phenomena from diverse perspectives.
By employing these diverse research methods and practices, our cohort of BRAID fellows have a unique opportunity to directly contribute to a richer and more nuanced understanding of how responsible AI interacts with human experiences, cultural dynamics, and the complexities of contemporary society, and the risks and benefits posed by developments in AI.
Will research approaches that are less fully within the AHRC discipline list/remit be considered provided they address the challenge? For example, qualitative research methods drawn from the social sciences?
A key aim of the BRAID programme is to embed voices and perspectives from the arts and humanities into the responsible AI ecosystem. However, the boundaries between the arts and humanities and fields like the social sciences can often be difficult to determine and we are adopting a flexible approach that embraces interdisciplinary methods and transdisciplinary knowledge. We recognise that the social sciences, along with other disciplines, have important contributions to make in many of the challenge contexts. Approaches and research methods drawn from outside of the arts and humanities are permitted where needed to meet the proposal’s objectives and the aims of the funding opportunity. What is important is that a significant portion of the research (i.e. at least 51%) is shown to be informed by or based in the arts and humanities.
Is Responsible AI considered in terms of the AI itself being responsible? Or its development? Or the responsible use of AI?
We are adopting a broad understanding of responsible AI as involving all the approaches mentioned.
Should projects be limited to challenges around general AI or a specific use case?
Projects need not be limited to general AI and specific use cases are acceptable, so long as the approach is well-justified in the proposal and contributes to moving the responsible research and innovation ecosystem forward.
Does the outcome/societal impact have to be uniquely only applicable to the UK or could it be further afield.
Societal impacts need not be geographically limited and we expect in many cases will not be.
I am wanting to work on a project in the Global South. Is the call focused on a UK context or are international collaborations eligible?
While you will need to be based at a UK research organisation, we welcome international collaborations with stakeholders from overseas, including from the Global South.
I would like to express interest in applying for a challenge hosted by a lab in Massachusetts. Is it possible to conduct this project remotely from the UK?
Absolutely. Given the lab for this challenge is based in the USA, both BRAID and the stakeholder expect much of the collaboration to conducted remotely. We recommend when developing your proposal that you consider the need to include a site visit or two to the overseas lab to support delivery of the project. Travel and accommodation costs will need to be factored into the budget proposal. To facilitate this, our Expression of Interest form provides an area for you to raise queries about site visits directly with the stakeholder. You’ll find this form in the Documentation section of our website here.
What characteristics would the non-academic stakeholders have? Do they need to be big companies or could they be smaller charities?
We use the term non-academic stakeholders (NAS) to refer to non-academic parties/organisations who have a demonstrable stake in the UK’s AI ecosystem. This can include organisations from the public, private or third sectors. For example, you might seek to collaborate with a large multinational technology company developing AI, a community group concerned about the effects of AI on their learners, or even a creative entity whose work is subject to use as data by an AI system.
I submitted to the BRAID Scoping to Embed Responsible AI in Context call and wish to apply to the BRAID Fellowships scheme. Please could you confirm that this is permissible?
Yes, you can apply to both the scoping and fellowships schemes. However, it should be noted that standard funding rules apply. Each proposal must be distinct and when taken together must not equate to more than 100% of your time. Otherwise, if you were successful in both applications, you would need to choose between them.
Is there an estimated number of academic-led projects that will be funded or does this depend on the number of challenge-led projects funded?
Projects will be awarded on the strength of the applications received, but as a starting point we are expecting an equal split between challenge-led and academic-led proposals. Our overarching aim is to create a balanced portfolio of projects, considering the diversity of teams, research themes, and geographies in that process.
Regarding the challenge-led pathway, are the non-academic stakeholders participating in the assessment process? Relatedly, am I correct in assuming that there is considerable scope to define projects within the challenges that the non-academic stakeholders have provided?
Yes, you are correct on both counts. The non-academic stakeholder will be involved in the assessment process and projects will be assessed on whether they address the challenge’s needs. However, there is considerable scope to define projects according to your own approach and methodology and we encourage a broad range of responses.
Are the fellowships something that could be done in conjunction with existing teaching (scaled back/part time maybe)? What salary grades are eligible?
Yes, the funding call is working with the standard UKRI model. Academics at all stages of their career are eligible and the university is responsible for costing your salary. We also appreciate that working full-time on the fellowship may not be possible in various circumstances that include teaching commitments. Where necessary, a fellow could apply at 40% FTE, for example, provided your research can be successfully completed within the time constraints.
I would like to apply but have a query about what would happen if I require a period of leave (for example, maternity leave) during the course of the fellowship. Is there any flexibility for people to still undertake research if leave is required at some point?
Yes, the funding call is working with the standard UKRI model. While grants are cash limited and expenditure against the grant must not exceed the value awarded, there is flexibility to cover parental leave, with provision for the fellowship to be placed in abeyance during the absence and extended by the period of absence.
Can I hire a research associate?
Unlike other grants, Fellowships are designed to be for individual researchers to undertake. While you can include the cost of a short period of research assistance or technical assistance to support specific activities in support of your research project, we expect the fellow to do most of the proposed research activity. We encourage you to refer to the principles of the Research Development Concordat and the AHRC guidance on training and developing early career researchers in the arts and humanities when planning and facilitating the work of any research assistants.
I am currently doing my PhD, will this fellowship available in the next academic year as well?
We are working towards there being a second round of fellowships and will have more information about this in 2024. Meanwhile, we are planning other BRAID programme activities over the next academic year and invite all responsible AI researchers to engage with us. Be sure to subscribe to our mailing list here to receive updates of future opportunities.
I have recently completed by PhD and don’t presently have a research affiliation. What would be required from my university to support my application?
In terms of the level of support offered by the university, the main thing that attaches to eligibility is that the university commits to employing you to undertake the project for the duration of the grant. Whether that is on a fixed-term or permanent basis does not affect eligibility. We also want to see why the organisation is supporting the project and its deliverables in terms of the benefits to your career, the field of responsible AI and any other relevant beneficiaries and stakeholders. Given you are an ECR we’d also be interested in information relating to how the organisation will support your career development and training. A template Letter of Support is available in the Documentation section of our website here.
I’m currently a postdoc research fellow and my contract will come to an end this year. Does it mean I’m not eligible to apply for the fellowship?
You do not need a contract with the research organisation to be eligible to apply. In the letter of support, the research organisation will need to commit to contracting you to work on the project for the duration of the fellowship if your application is successful.
Do you have to be working in a university in order to apply for the challenges?
You do not need to be currently working at a university to apply. However, to be eligible your proposal will need the support of UK research organisation and that support must include a commitment to employ you to work on the project for the duration of the fellowship in the event of a successful application.
I am a researcher and arts practitioner. Are standard research costs included in the scope of funding?
Yes, standard research costs are included. Costings need to be comprehensive and justifiable in terms of delivering on the project’s objectives, contributing to your development as a researcher/arts practitioner in the field, and contributing to our goal of embedding arts and humanities research into the responsible AI ecosystem.
Should the budget proposal include the research overhead rates of the university?
Research overhead rates are only applicable for the mentor’s costs which is being funded at 80% fEC and includes salary, estates and indirects. Otherwise, only direct costs for the fellow’s salary, travel and research-related expenditure are eligible at 100%.