Draft: The Multipolar Silicon: Why the BRICS AI Declaration is a Sovereign Pivot for the Global South

Summary Glossary

• The Shift: BRICS science academies move from AI safety rhetoric to AI as core scientific infrastructure.
• Compute Sovereignty: A direct call for shared high-performance computing to break the Western hyperscaler monopoly.
• The Multilingual Gap: Emphasis on local-language LLMs to prevent cultural and data colonialism.
• Scientific Temper 2.0: Redirecting AI from consumer chatbots to materials science, drug discovery, and climate modeling.

The global conversation on Artificial Intelligence has, until now, been largely dictated by a binary of Silicon Valley's commercial optimism and the European Union’s regulatory caution. However, a significant third pole just solidified in New Delhi. Under India’s BRICS Presidency, the Indian National Science Academy (INSA) convened the BRICS Science Academies Forum, drafting a declaration that fundamentally reframes AI not as a risk to be managed, but as a sovereign instrument for scientific discovery.

This isn't just another diplomatic communiqué. It is a tactical blueprint for what we might call 'Scientific Multipolarity.'

Breaking the Compute Moat

The most critical takeaway from the forum—attended by Brazil, China, Russia, South Africa, and newer members like Egypt and Ethiopia—is the consensus on shared computing infrastructure. For too long, the 'Digital Divide' has been defined by bandwidth; in the AI era, it is defined by FLOPS (Floating Point Operations Per Second). By proposing a collaborative data and compute framework, the BRICS nations are effectively signaling their intent to bypass the "toll-booths" of Western cloud giants.

If successful, this allows a researcher in Addis Ababa or Hyderabad to access the same high-tier compute power required for complex climate modeling without being tethered to a proprietary, dollar-denominated subscription.

From Chatbots to Cold Fusion

There is a refreshing absence of 'AGI' hype in the INSA-led discussions. Instead, the focus is squarely on "AI for Science." The declaration prioritizes materials science, drug development, and engineering. This is where India’s "Scientific Temper"—the constitutional duty to develop inquiry and humanism—finds its digital expression. When we use AI to decode the molecular structure of a new superconductor rather than generating generic marketing copy, we are moving from being consumers of technology to architects of it.

The Multilingual Front

The call for multilingual AI models is perhaps the most "Human-First" aspect of the declaration. Current LLMs are notoriously Anglo-centric, often hallucinating or failing when dealing with the linguistic nuances of the Global South. By pushing for models that respect local context and data, the BRICS science academies are building a firewall against 'Data Colonialism'—the process where Global South data is extracted to train models that are then sold back to them.

The BharatLens Deduction

The upcoming in-person meeting at IIT Hyderabad in July won't just be about signing a paper. It will be the start of a "Compute-Non-Aligned Movement." Just as India led the way in sovereign digital public infrastructure (DPI) with UPI, it is now positioning itself to lead the world in Sovereign Scientific AI. The message is clear: The future of intelligence will not be centralized; it will be distributed, multilingual, and, most importantly, sovereign.

Sources

• Press Information Bureau: INSA convenes BRICS Science Academies Forum on AI for Sustainable Development. [https://www.pib.gov.in/PressReleasePage.aspx?PRID=2270686]
• Tribune India: Global South Roadmap on Responsible AI. [https://www.tribuneindia.com/news/business/insa-convenes-brics-science-academies-forum-on-ai-for-sustainable-development-focuses-on-inclusive-global-cooperation/]
• BioSpectrum: 10 Countries draw up AI roadmap for Science. [https://www.biospectrumindia.com/news/22/27926/10-countries-come-together-to-draw-up-roadmap-for-global-south-on-responsible-ai.html]