# Power to the People? Opportunities and Challenges for Participatory AI

**Authors:** Abeba Birhane, William Isaac, Vinodkumar Prabhakaran, Mark Díaz, Madeleine Clare Elish, Iason Gabriel, Shakir Mohamed  
**arXiv:** [2209.07572](https://arxiv.org/abs/2209.07572)  
**Date:** 2022-09-15  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2209.07572v1/

## Key points

- Growing momentum of participatory approaches to AI and ML
- Participation seen as gateway to inclusive, equitable, robust, responsible, and trustworthy AI
- Crucial for understanding and representing needs, desires, and perspectives of historically marginalized communities
- Review of participatory approaches within historical contexts
- Examination of participatory methods and practices in the AI and ML pipeline
- Presentation of three case studies demonstrating potential for beneficial, emancipatory, and empowering technology design, development, and deployment
- Importance of developing contextual and nuanced understanding of participation
- Consideration of primary beneficiaries crucial for realizing benefits and opportunities of participation
- Specific examples related to Māori data sovereignty laws requiring key decision-making by Māori communities for downstream applications using their language data
- Emphasis on alignment with Māori values when utilizing their data
- Need to address limitations and concerns while ensuring comprehensive understanding of participation's scope

## Summary

The paper discusses the growing momentum of participatory approaches to artificial intelligence (AI) and machine learning (ML), emphasizing that participation is seen as a gateway to inclusive, equitable, robust, responsible, and trustworthy AI. These approaches are crucial for understanding and representing the needs, desires, and perspectives of historically marginalized communities. The authors first review participatory approaches within historical contexts and examine participatory methods and practices in the AI and ML pipeline. They then present three case studies in participatory AI to demonstrate its potential for beneficial, emancipatory, and empowering technology design, development, and deployment. The paper highlights the importance of developing a contextual and nuanced understanding of participation. It argues that considering who the primary beneficiaries of participatory activities should be is crucial for realizing the benefits and opportunities that participation brings. Moreover, it delves into specific examples related to Māori data sovereignty laws requiring key decision-making by Māori communities for downstream applications using their language data. The Māori community believes that any project utilizing their data must align with their values. Overall, the paper emphasizes that while participatory AI holds great potential, it is essential to address limitations and concerns while ensuring a comprehensive understanding of participation's scope. By doing so we can harness the benefits of participation while avoiding potential pitfalls.