Collectively Reimagining Artificial Intelligence With Marginalized Communities

Paper Title

Collectively Reimagining Artificial Intelligence With Marginalized Communities

Publication Info

• Topic area: Speculative co-design of AI with marginalized communities.
• Keywords: Artificial Intelligence, marginalized communities, co-design, speculative design, cultural preservation, ethical AI, socio-economic advancement, Ocean View, Afrikaaps, community-centered AI.

Background and Problem

• Problem / challenge: Marginalized communities, especially in Africa, are underrepresented in AI design and deployment, leading to cultural erasure, exclusion from technological benefits, and reinforcement of socio-economic inequalities.
• Significance: Addressing these gaps is critical for ensuring equitable access to AI technologies and fostering socio-economic development while preserving cultural identity.
• Motivation and related work: Prior research has explored participatory and speculative design approaches to democratize technology development, but there is limited engagement with marginalized communities to elicit their visions for AI futures. This study builds on these approaches to address the gap.

Solution

• Proposed approach: Speculative co-design workshops with Ocean View residents to collectively reimagine AI futures tailored to their cultural, social, and economic contexts.
• Novelty:
  1. Integration of local language (Afrikaaps) and cultural identity into AI designs.
  2. Exploration of affordable, offline-capable AI systems for resource-constrained environments.
  3. Community-driven visions for leveraging AI to support socio-economic advancement.
  4. Use of artistic impressions to facilitate engagement and imagination in co-design processes.
• Procedure and key techniques:
  • Conducted three workshops focusing on AI readiness, speculative design practices, and community-centered AI futures.
  • Participants engaged in brainstorming, prototyping, and envisioning AI artefacts tailored to their needs.
  • Artistic impressions of AI were used to demystify technology and spark creative dialogue.
  • Data collected through field notes, audio recordings, and thematic analysis.

Results

• Concrete findings:
  • Participants envisioned AI systems that preserve cultural heritage, such as Afrikaaps voice automation and educational tools.
  • Artefacts included AlvinBot (AI house for community services), Dinges (universal language translator), and prototypes for sustainable farming and fitness assistance.
  • Identified community priorities such as affordable AI, offline functionality, and socio-economic mobility support.
• Advantage over baselines:
  • Centered marginalized voices in AI design, addressing cultural erasure and socio-economic exclusion.
  • Proposed locally relevant and affordable AI solutions, contrasting with mainstream AI optimized for resource-rich environments.
• Experiments / evaluation:
  • Workshops involved 24 participants, with iterative engagement across three sessions.
  • Participants co-designed speculative artefacts reflecting their lived experiences and aspirations.
  • Data analyzed through reflexive thematic analysis to identify themes and insights.
• Limitations and future work:
  • Limited participant diversity; future studies should engage broader community segments.
  • Attrition across workshops due to caregiving and work schedules.
  • Need for open-ended prompts to explore non-AI-centric solutions.
  • Future work includes developing Afrikaaps language models and expanding AI literacy programs.

Summary

This study explored speculative co-design with Ocean View, a marginalized community in South Africa, to reimagine AI futures that preserve cultural identity, address affordability, and support socio-economic advancement. Workshops generated artefacts such as AlvinBot and Dinges, reflecting community priorities for accessible and culturally grounded AI systems. Findings highlight the need for ethical, inclusive AI designs that align with the lived realities of marginalized groups. Future work involves implementing these ideas and expanding AI literacy to ensure equitable access and meaningful engagement.