Participatory AI for Humanitarian Innovation: a Briefing Paper
Authors: Aleks Berditchevskaia, Kathy Peach, Eirini Malliaraki
A Nesta briefing paper outlines approaches to the participatory design of AI systems, and explores how these approaches may be adapted to design collective crisis intelligence solutions for humanitarian settings
CCI combines artificial intelligence (AI) with methods that gather on-the-ground human intelligence from communities affected by crises and frontline responders.
The paper accompanies Nesta’s report Collective Crisis Intelligence for Frontline Humanitarian Response, which analyses how CCI is being used to improve anticipation, management and response in the humanitarian sector.
Participatory AI or participatory machine learning in their broadest sense refer to the involvement of a wider range of stakeholders than just technology developers in the creation of AI systems, models, tools or applications.
Participatory AI for Humanitarian Innovation: Aa Briefing Paper responds to the growing interest in using participatory approaches in designing, developing and evaluating AI systems across industry, academia and the public sector.
Based on a rapid analysis of existing participatory AI case studies and the academic literature, the paper proposes a conceptual framework for such participatory approaches.
The framework outlines participatory interventions that could be used at different stages of AI design and development.
The methods involved fall into four categories:
- Consultation – refers to participation where input occurs outside the core AI design and development process, and it is not guaranteed that it will impact the design of AI. Common research methods include focus groups and interviews, surveys and deliberative approaches.
- Contribution – refers to participation that is usually time-limited to one stage of the AI design and development processpipeline. External stakeholders complete one of the tasks that is necessary to AI development (e.g. data collection, data labelling, validation of model outputs). Common methods include both targeted and open crowdsourcing.
- Collaboration – refers to participatory practices with multiple touchpoints during along the design and development process, pipeline and/or where external stakeholders can meaningfully contribute to interrogating the model and shaping the features that it uses to make predictions or classifications, even if they were not involved in problem setting.
- Co-creation – is the most comprehensive form of stakeholder involvement in AI design and development. It involves engagement at multiple stages throughout the processpipeline, as stakeholders discuss their needs, values and priorities.
The report illustrates the framework using three in-depth case studies in the humanitarian, development and other sectors.
It also suggests five key design questions to help with designing participatory AI projects:
- Who defines the process and what counts as success?
- Whose participation is required?
- What is the intent behind participation?
- How will participants be rewarded?
- What is the process for closing the project?
Applying participatory AI approaches in humanitarian settings
The final section of the report explores the relevance of participatory AI to the humanitarian sector. Drawing on the Core Humanitarian Standard, the report highlights how AI systems may jeopardise humanitarian principles or the rights of communities affected by crises, and gives examples of how participatory approaches could mitigate some of the risks.
Although participatory design alone is not enough to address all critiques of AI in humanitarian settings, when developed alongside complementary measures it could strengthen the ecosystem for responsible AI.
The briefing paper informs Nesta’s approach for testing and evaluating participatory AI in humanitarian contexts as part of a larger project on CCI for humanitarian action in partnership with the IFRC Solferino Academy, with additional support from the Alan Turing Institute and the Digital Civics Centre (led by Open Lab at Newcastle University).
The project was funded by a grant from the UK Humanitarian Innovation Hub (UKHIH). UKHIH is an independent entity hosted within Elrha and fully funded by the UK Foreign, Commonwealth & Development Office.