*The risk that ‘large-scale’ AI won’t serve local communities is real!*
Excitement over generative AI and ever-larger models, typically delivered through networked platform infrastructures, has overshadowed the importance of building systems for specific communities and local needs. That is, AI systems built for the “majority” often fail minority communities. With 88% of the global population living outside of the West, many systems transplanted will not work effectively and may harm the many diverse local communities within a country or across the “global south”. Further, as states debate regulation and companies offer "assurance", these are often far from or exclude minority communities, context, needs and other considerations.
Towards this, we will convene technologists, researchers, civil society and others advocating for and building AI-related technologies for local communities. The goal is to share experiences, foster dialogue, and generate a corpus of lessons learned - from different perspectives, approaches, contexts and geographies - to explore what works and what doesn’t in designing AI systems for and with local communities and the roles of networked platforms within this space.
Workshop will include real-world stories, interactive exercises, and a "learnings from practice" fireside chat, with narratives from Brazil, Argentina, Oceania, South Africa and Egypt, by people who have built AI driven systems for their local community. We will facilitate exercises and discussion that gets attendees to consider approaches they would take regarding tech, considering specific roles, needs, barriers, and tensions (e.g. resources, power, system dependencies, political economy, etc). It culminates with insights from different engagements, providing approaches and understandings on how to “build for the long tail” – the many different local contexts and consideration of tensions in value sensitive design.
The workshop aims to “rupture” the dominance of the networked platforms of big tech and dominant Western views. We will do this by spurring collaboration, insights and perspectives across the range of participants to (a) re-think AI development and digital governance for local communities, (b) provide practical ways forward on approaches for understanding and building human-centred technology in and for local communities, (c) encourage coalitions including diverse voices in small, local contexts in AI and digital platform research, development and governance.
Intended schedule of half-day workshop is as follows:
Introduction: Explain who we are, why we are here and what to expect
Session 1 (Interactive Storytelling & Surveying):
There will be 3 to 4 stories. For each story, encompassing a real-world situation, storytellers will first provide a brief synopsis of the situation in which digital technology was affecting their local communities. After, attendees will be surveyed on what they would do in that situation and why. Storytellers will then share what they did and why, bringing to light successes and challenges. Attendees will be re-polled to learn if they would change their approach, after having heard storytellers’ journeys.
Situations have inherent tensions between local values and Western views that may or may not be explicit, and may also be interpreted differently according to one’s sociocultural background, value systems, etc.
Session 2 (Facilitated Discussion):
We will have a facilitated discussion, identifying opportunities and challenges to building locally. In small groups, attendees will explore what is needed to build locally, rupturing the current ecosystem and supply chain of big tech and dominant Western views. A representative will report back to the larger group in an interactive discussion.
Exercises will help surface and interrogate discussions in an interactive participatory environment.
Wrap-up: Facilitators will summarise key takeaways and opportunities to rupture the status quo, fostering approaches for AoIR community, in collaborating with and building sustainable AI and supporting digital technologies for local communities.
Broader aim of this workshop is to produce a paper, summarise findings and learnings that will be helpful for researcher and practitioner communities.
Note: the people "checked" as speakers are the main organisers of the workshop.