The policy briefing “Fostering a Federated AI Commons Ecosystem”, by Joana Varon, Sasha Costanza-Chock and Timnit Gebru, makes the case plainly. The development of Artificial Intelligence (AI) is currently spearheaded by a handful of Big Tech firms based in the global minority, racing to outpace each other and turbocharging surveillance capitalism, digital colonialism, and a monoculture of thought. As a result, most AI systems are being developed and deployed with a “one-model-for-everything” approach that increases inequities, automates oppression, and exacerbates the climate catastrophe.

This policy briefing, submitted to T20, the G20 engagement group of think tanks and research centers, provides actionable recommendations for the G20 to foster decentralized AI development.

We urge support for an alternative AI ecosystem characterized by community and public control of consensual data; decentralized, local, and federated development of small, task-specific AI models; worker cooperatives for appropriately compensated and dignified data labeling and content moderation work, and ongoing attention to minimizing the ecological footprint and the social-economic-environmental harms of AI systems.

This brief was submitted to the T20 Task Force 05, on “Inclusive digital transformation”, under the subtopic “Challenges, Opportunities, and Governance of Artificial Intelligence”.

Download the T20 POLICY BRIEFING: Fostering a Federated AI Commons ecosystem here.

Abstract

The development of Artificial Intelligence (AI) is currently spearheaded by a handful of Big Tech firms based in the global minority, racing to outpace each other and turbocharging
surveillance capitalism, digital colonialism, and a monoculture of thought. As a result, most AI systems are being developed and deployed with a “one-model-for-everything” approach that increases inequities, automates oppression, and exacerbates the climate catastrophe. This is counter to the Agenda 2030 goals of “promoting the social, economic and political inclusion of all,” and in direct opposition to the three pillars of sustainable development – social, economic and environmental – as defined in Rio-92 and reaffirmed at Rio+20. If the development of AI systems fails to acknowledge and redress structural inequities, these systems will continue to cause more harm to marginalized communities and territories.

However, the centralization of power through AI is not inevitable. For example, there are initiatives aiming to build federations of small organizations that can become part of a broader AI Commons ecosystem. This policy paper provides actionable recommendations for the G20 to foster decentralized AI development. We urge support for an alternative AI
ecosystem characterized by community and public control of consensual data; decentralized, local, and federated development of small, task-specific AI models; worker
cooperatives for appropriately compensated and dignified data labeling and content moderation work, and ongoing attention to minimizing the ecological footprint and the
social-economic-environmental harms of AI systems. We call on the G20 to center bienes comunes (the commons), human rights, and the public’s interest in AI development.

Keywords: AI, Big Tech monopolies, decentralization, federated AI, distributed AI, bienes comunes, commons, social-environmental justice