A methodological stance for participatory AI
AI from the Margins
A methodological stance that centers the lived experiences of minoritized communities before participatory AI design begins — ensuring their standpoint shapes what AI is for, before problem definitions, success criteria, or design artifacts are set.
About
Why it matters
AI is increasingly embedded in healthcare — shaping clinical decision-making, disease detection, and support for daily living. Despite its promise, AI applications can reproduce and amplify existing inequities, creating particular vulnerabilities for minoritized communities.
Participatory AI promises to involve affected communities in shaping the systems that affect them. In practice, however, participation typically starts only after problem definitions and success criteria have already been set — leaving limited room for communities to reshape what an AI system is actually for.
AIM addresses this gap. Rather than treating lived experience as one input among many within an existing design process, AIM positions it as the preparatory foundation from which participatory AI design begins. This is not a procedural adjustment — it is an epistemic one: the standpoint of those most affected by AI's harms is the starting point, not an afterthought.
This work continues beyond publication. As participants themselves urged, the stories shared and lessons learned should feed change in healthcare and policy, not only research about them.
The stance
Seven preconditions
These preconditions are mutually constitutive.
They cannot be implemented independently — each one creates the conditions under which the others can function.
They cannot be implemented independently — each one creates the conditions under which the others can function.
Publications
Research
2026
AI From the Margins (AIM): Rethinking Participatory AI Design Through the Lived Experience of Minoritized Communities
The full paper introducing AIM as a methodological stance, presenting its seven preconditions, and reporting findings from four lived experience sessions with thirteen women and non-binary people of color in a Dutch healthcare context.
Accepted — AIES 20262026
BNIM in participatory AI design: centering minoritized lived experiences in healthcare
Earlier work applying the Biographical Narrative Interpretive Method (BNIM) in healthcare AI research, surfacing requirements for AI grounded in participants' own accounts. The methodological foundation that AIM builds on.
ACM FAccT 20262025
Decoding AI ethics: What do ethical tools tell us about ethics?
Our findings add empirical support to recent AI and ethics research advocating for bottom-up approaches that embrace diverse perspectives, challenging the dominance of broadly stated ethical principles, diagnostic tools, and an overreliance on self-assessments.
Ethics and Information TechnologyIn practice
One application
AIM's preconditions were enacted in one application, through four sessions in a Dutch healthcare context. The techniques below are substitutable — the preconditions are not.
01
Sharing lived experience
Each participant tells their story in their own words, prompted by a Single Question aimed at Inducing Narrative (SQUIN). Peers ask follow-up questions through Socratic dialogue. Reciprocity and co-ownership are established from the outset.
May span multiple meetings to ensure every voice is heard fully.02
Collective rules
Individual narratives become the foundation for collective reflection. Participants articulate shared rules for equitable practice — moving from personal accounts toward principles they define and own. A decolonizing move: from objects of study to agents of inquiry.
03
If, where, and how
Participants determine whether AI should play a role at all, and if so where and how — including through visual methods such as Rich Picture. Technical framing enters only after lived experience has been articulated on its own terms. Refusal of AI is a legitimate outcome.
04
Encountering policy
Policy workers join participants to discuss how AI governance might safeguard lived experiences. The session surfaces — rather than smooths over — asymmetry between institutional and experiential framings.
In our application, that asymmetry proved harder to bridge than anticipated. What institutions do after such an encounter is the question we are pursuing next.Theoretical grounding
Built on the work of
AIM draws on and synthesizes a body of scholarship across standpoint theory, decolonizing research, feminist design, and participatory methods.
- Costanza-Chock, S. (2020). Design Justice: Community-Led Practices to Build the Worlds We Need. MIT Press.
- Collins, P. H. (2000). Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. Routledge.
- Smith, L. T. (2012). Decolonizing Methodologies: Research and Indigenous Peoples. Zed Books.
- hooks, b. (1994). Teaching to Transgress: Education as the Practice of Freedom. Routledge.
- Crenshaw, K. (1991). Mapping the margins: intersectionality, identity politics, and violence against women of color. Stanford Law Review, 43(6).
- Wengraf, T. (2001). Qualitative Research Interviewing: Biographic Narrative and Semi-Structured Methods. SAGE.
- Sloane, M. et al. (2022). Participation is not a design fix for machine learning. ACM EAAMO '22.
- Tuck, E., & Yang, K. W. (2014). Unbecoming claims: pedagogies of refusal in qualitative research. Qualitative Inquiry, 20(6).
Contact
Get in touch
Interested in AIM, a collaboration, or a presentation? Reach out directly below.
Email directly at portc@hr.nl
In collaboration with
