Title: AI From the Margins (AIM): Reimagining Participatory AI Design via the Lived Experiences of Minoritized Groups

Abstract: Artificial intelligence systems often perpetuate and exacerbate the structural disparities that minoritized communities endure. While participatory AI has emerged as a potential remedy, conventional approaches typically engage stakeholders only after the core problems and success metrics have already been established. This late-stage involvement restricts the ability of minoritized groups to fundamentally redefine the purpose of AI technologies. To address this gap, we introduce AI From the Margins (AIM), a methodological framework designed to identify the specific conditions necessary to elicit, center, and sustain the lived experiences of these communities within participatory AI design processes. Rather than functioning as a rigid protocol, AIM outlines a set of preconditions that can be implemented through various techniques across diverse contexts. We tested this approach in the Netherlands through eight sessions involving 13 women and non-binary people of color alongside five municipal policy officials. The methodology employed four key strategies: (1) using the Biographic Narrative Interpretive Method (BNIM) for narrative elicitation; (2) establishing co-constructed rules; (3) allowing participants to decide the extent, location, and manner of AI involvement; and (4) facilitating dialogue to translate personal experiences into AI policy recommendations. Post-session feedback indicated that participants viewed the engagement as meaningful and advocated for its expansion. These results highlight how a foundational commitment to lived experience can fundamentally reshape the objectives and outcomes of participatory AI design.

Source: arXiv Generated at: 2026-06-02 00:00:00 UTC