The ColibrAI research project aims to lay the foundation for a long-term research agenda on community-orientated AI by fostering interdisciplinary collaboration and cultivating a culture of critical engagement in AI development.

The project aims to challenge prevailing narratives on AI that often reinforce power imbalances by focusing on practical measures that empower marginalised groups to actively co-create AI policies, practices and governance.

At its core, ColibrAI strives for a participatory AI framework that integrates the principles of justice, equality and shared liberation into AI research and education. Rather than allowing AI development to be dictated by centralised, corporate or state-driven power structures, the project explores ways in which communities can meaningfully participate in the development, testing and governance of AI systems.
This requires a rethinking of AI as not just a neutral technology, but as a social and political tool that needs to be designed and deployed in a transparent, accountable way that is consistent with democratic values.

With the help of Unite! University, ColibrAI will initiate a collaboration between researchers and institutions to

• Promote interdisciplinary knowledge exchange by pooling expertise in AI, participatory design and decolonial perspectives to explore alternative AI frameworks.
• Investigate ethical AI governance models, examining how decentralised, community-based decision-making processes can replace top-down AI policy structures.
• Strengthening the "Unite!" research network by fostering new collaborations between KTH, Aalto and TUDa, integrating technical and social science perspectives on AI. As the ColibrAI project represents the first collaboration between its PIs, the start-up funding from Unite! plays a crucial role in building a sustainable interdisciplinary network.
• Empowering researchers, students and practitioners by providing tools to critically analyse and redesign AI systems through education, workshops and collaborative research.
• Laying the groundwork for long-term research and funding to position ColibrAI for larger initiatives beyond this initial start-up phase.

With a community-orientated and equity-focused approach, ColibrAI aims to fundamentally rethink how AI is developed, who can participate in its design and the implications of this.

The aim is not only to criticise existing AI paradigms, but to actively create new, participatory models that combat exploitation and democratise the governance of AI.

Existing AI initiatives within Unite! primarily emphasise technological innovation and interdisciplinary research, but do not explicitly focus on participatory AI frameworks, decolonial perspectives or community-oriented governance models.

This makes ColibrAI a strategic and timely initiative that can bridge this gap by bringing together KTH (Sweden), Aalto (Finland) and TUDa (Germany) in an interdisciplinary research collaboration. Each partner university brings unique expertise, providing a strong foundation for research in participatory AI:

• KTH brings expertise in AI justice and data feminism, focusing on bias, fairness and power structures in AI systems.
• Aalto contributes knowledge on participatory design and community-orientated creative practices that question the instrumental use of AI and offer methods for community-orientated technology development and governance.
• TUDa offers a decolonial perspective on AI, digital technology and pedagogy, ensuring that the research challenges dominant westernised AI narratives and gives a central role to marginalised voices.

By utilising this combination of technical, creative and social expertise, ColibrAI will build a novel interdisciplinary research network that critically engages with the societal implications of AI and creates alternative frameworks for equitable and community-oriented AI design and development.

Phase 1: Establishing collaboration and knowledge exchange

• Organising online meetings to align expectations, define common research goals and capture partners' expertise.
• Conduct literature and initiative reviews to assess existing community-led AI research and identify gaps, which will be published via a collaborative and public platform.
• Conduct online knowledge sharing session
• Produce public research gap reports summarising key findings from the literature review to provide a common basis for future work.
• Defining a preliminary funding plan, identifying key funding calls and agreeing Research Areas for later phases.
• Establish a joint research archive of results and references to facilitate collaboration (the structure of the archive will be determined based on the needs of the partners).

Phase 2: Developing the research agenda and engaging the public

• Integration of discussions on participatory AI, shared liberation and ethical governance of AI into the partner universities' courses through guest lectures and online discussions
• Organising a one-day online workshop on participatory AI, workers' rights and decentralised decision-making involving researchers, civil society groups, policy makers and industry representatives.
• Identify potential joint research projects and interdisciplinary synergies.
• Start developing initial research concepts involving knowledge exchange activities and feedback from stakeholders.

Phase 3: Face-to-face workshop and research planning

• Organise a two-day face-to-face workshop at KTH, bringing together researchers, PhD students and external experts to refine research directions and strengthen collaboration.
• Completion of at least two joint research concepts, which are to be developed into larger funding applications.
• Produce small joint publications (e.g. blog posts, essays or short research papers) to increase the visibility of ColibrAI and achieve broader participation
• Production of a white paper on participatory governance of AI, summarising the results of the workshop and research directions for policy and dissemination of research results

Phase 4: Elaboration of the proposal and future funding strategy

• Identify and evaluate potential funding opportunities, focussing on Horizon Europe, ERC, MSCA and NordForsk funding.
• Drafting of joint research proposals and clear division of writing tasks among the partners.
• Develop a risk mitigation plan and identify alternative funding sources (e.g. small grants, industry partnerships or pilot projects).
• Finalise a timeline for the continuation of ColibrAI to ensure the long-term sustainability of the research beyond the seed funding phase.

Expected results:

• Establishment of a research network on participatory AI, promotion of interdisciplinary collaboration between KTH, Aalto and TUDa.
• Engage researchers, students and community members through online sessions, workshops and teaching events.
• Developing a formal research conceptualisation that integrates AI, participatory design and decolonial perspectives.
• Produce blog posts and white papers on participatory AI governance to inform future research and policy discussions.
• Finalising joint research concepts for large-scale funding applications.
• Strengthen the research capacity of Unite! in the field of community-orientated AI by building new academic and policy collaborations.
• Positioning the project for future large-scale funding to ensure its long-term sustainability.