As part of the India AI Impact Summit in New Delhi, we participated in the roundtable AI Governance from the Global South: From Redlines to Structural Baselines, organized by IT for Change alongside a broad network of partners. This space brought together civil society organizations, researchers, and public policy actors from across the Majority World, with a clear objective: to move AI governance debates beyond identifying risks or harmful use cases, and to challenge the structural foundations on which the current technological paradigm is built. Rather than focusing solely on mitigating harms once systems are already deployed, the conversation centered on defining non-negotiable conditions across the entire AI value chain.

In other words, what needs to change at the root for these systems to be compatible with human rights, social justice, and ecological sustainability? Building on this premise, the dialogue was structured around key questions:
‍
*What governance baselines must be in place for AI systems to align with principles of justice and rights?*What would it mean to treat labor regulation, public procurement, competition policy, and environmental oversight as central components of AI governance?
*How can governance reshape the political economy of AI innovation, rather than merely respond to its downstream impacts?

These discussions reflect a shared concern: the current model of AI development is deeply shaped by power concentration, labor precarity, exclusion, and environmental harm. In response, it is not enough to regulate we must reimagine the paradigm. In this context, the roundtable engaged with the framework ReGenAI: A New Deal for the AI Economy, an alternative proposal that calls for reorienting AI toward more diverse economies, strengthened local capacities, democratic governance, public value, and ecological sustainability. This New Deal is grounded in shared political and ethical commitments, and advances a vision centered on:

*Meaningful and dignified work
*Diversified economies
*Pluralistic knowledge societies
*Planetary flourishing
‍
More than a statement, it is an invitation to actively contest and reshape the trajectory of artificial intelligence at a global level. Being part of this space reaffirms something we strongly believe at Diversa: it is not enough to make AI less harmful. We must transform the structural conditions that define how it is designed, who benefits from it, and who is left out. We extend our deep gratitude to IT for Change, the Global Digital Justice Forum (including organizations such as Data Privacy Brasil, Derechos Digitales, EngageMedia, ETC Group, Research ICT Africa, and Tech Global Institute), as well as the Ada Lovelace Institute, Centre for Communication Governance, Planetary AI Network, The Future Society, UNESCO, and Friedrich-Ebert-Stiftung for making this space possible.

We continue pushing for AI that is just, democratic, and grounded in context.