BRICS leaders call for stronger international cooperation, wider access to AI resources
The New Delhi Declaration 2026 dedicates a significant section to artificial intelligence, describing it as an opportunity to stimulate economic growth and sustainable development for member states
BRICS leaders called for enhanced international cooperation to widen access to AI resources — including compute, data, and talent — while ensuring AI remains safe, secure, inclusive, and reliable
The declaration commits members to implementing the BRICS Leaders' Statement on the Global Governance of Artificial Intelligence
Member states endorsed cooperation on safe, ethical, and human-centred use of AI in education, alongside digital literacy, academic mobility, and institutional partnerships
Leaders also recognized the importance of protecting intellectual property used in AI training and ensuring fair remuneration to rights holders
BRICS Leaders' Statement on the Global Governance of Artificial Intelligence (2025)
BRICS adopted its first stand-alone, high-level statement on AI governance at the 17th Summit in Rio de Janeiro on 6 July 2025. It set out guidelines for the responsible, non-military development and deployment of AI, emphasized that the United Nations should be the central, inclusive forum for global AI governance, and flagged data governance as essential for developing countries to benefit from the digital economy.
Key Details
- Adopted 6 July 2025 at the 17th BRICS Summit, Rio de Janeiro — the bloc's first dedicated AI governance declaration
- Scope explicitly limited to non-military AI applications
- Calls for AI governance to be anchored in the UN system rather than a narrower plurilateral body
- Followed by the Kazan Summit (2024) framework programme for AI cooperation, which had already launched the BRICS AI Research and Innovation Centre working on multilingual NLP, smart agriculture, and climate forecasting
The 2026 New Delhi Declaration explicitly commits members to implementing the 2025 Rio statement, showing a two-year progression from principles (2025) to an access/resource-sharing agenda (2026) within the bloc's AI cooperation track.
The "AI Divide" and Access to Compute
A central theme of the New Delhi Declaration's AI language is the global disparity in access to AI compute, data, and foundation-model capability between advanced and developing economies — often termed the "AI divide" or "compute divide." This has become a recurring theme at multilateral fora (G20, UN, GPAI) alongside BRICS, reflecting concern that AI benefits could concentrate further global inequality unless deliberately broadened.
Key Details
- Compute (GPU/data-centre capacity), data, and skilled talent are the three resource inputs most cited in "AI divide" discussions
- India's own domestic response is the IndiaAI Mission, approved by the Union Cabinet in March 2024 with an outlay of roughly ₹10,372 crore, built around seven pillars: Compute, Foundation Models, AIKosh (data platform), Application Development, FutureSkills, Startup Financing, and Safe & Trusted AI
- Under the IndiaAI Mission's Compute pillar, GPU capacity empanelled for public/researcher access had grown to over 34,000 GPUs by 2025, well past the original 10,000-GPU target
- The Mission also established the IndiaAI Safety Institute under its "Safe and Trusted AI" pillar to address AI risk and safety concerns
BRICS's call for "wider access to AI resources" mirrors, at a multilateral level, the same access constraints that India's domestic IndiaAI Mission was designed to address — situating India's own AI policy within the broader Global South narrative the declaration articulates.
AI, Intellectual Property, and Fair Remuneration
The declaration's reference to intellectual property rights in AI training data reflects an unresolved global policy question: whether using copyrighted material to train AI models constitutes fair use/fair dealing or requires licensing and remuneration to original rights holders. This debate spans copyright law frameworks in multiple jurisdictions and has no settled multilateral resolution yet.
Key Details
- India's copyright framework is the Copyright Act, 1957 (as amended), administered under the Ministry of Commerce and Industry; it does not yet contain AI-specific training-data provisions
- WIPO (World Intellectual Property Organization) has been convening member-state discussions on AI and IP since 2019 without a binding treaty outcome
- The EU AI Act (2024) and various national proposals require varying degrees of transparency on training-data sources, a comparative reference point for India's evolving policy
By naming IP protection and "fair remuneration to right holders" for AI training data, the BRICS declaration signals member states are beginning to converge on a rights-holder-protective stance, distinguishing it from more permissive fair-use approaches seen elsewhere.
- BRICS Leaders' Statement on the Global Governance of AI adopted 6 July 2025, Rio de Janeiro (17th Summit)
- BRICS AI framework cooperation programme dates to the 2024 Kazan Summit (16th Summit)
- IndiaAI Mission: approved March 2024, outlay approximately ₹10,372 crore, seven pillars
- IndiaAI Mission Compute pillar: exceeded 34,000 empanelled GPUs by 2025 against an initial target of 10,000
- BRICS now comprises 11 full members representing about 40% of global GDP
- New Delhi Declaration 2026 was adopted at the 18th BRICS Summit under India's chairship