Delhi Declaration Sets a New Global Course for AI Governance
The Delhi Declaration positions the Global South at the forefront of AI governance, blending innovation, equity, and sustainability into one framework.
As leaders from across the world assemble at Bharat Mandapam in New Delhi, February 19 marks what many are calling a watershed moment in global technology policy. The formal adoption of the Delhi Declaration at the AI Impact Summit 2026 is being described as the “Magna Carta of AI,” signalling a decisive shift in how artificial intelligence will be governed worldwide.
Unlike previous gatherings in Bletchley Park or Seoul that centred largely on the existential risks posed by frontier AI, the New Delhi framework takes a broader and more development-oriented view. It is the first major global AI governance blueprint to emerge from the Global South, placing equity, inclusion and growth at its core rather than focusing solely on regulatory caution.
The ‘Seven Sutras’ of Governance
Set to be unveiled during the Prime Minister’s keynote address at 10.25 am, the declaration is anchored in what policymakers describe as a “techno-legal” approach. Instead of rigid, compliance-heavy legislation that could slow innovation, it outlines seven flexible guiding principles — or Sutras — designed to evolve with the technology.
The principles include:
- Trust as the Foundation: Ensuring AI systems are reliable and secure.
- People First: Prioritising human agency and dignity in every algorithm.
- Innovation over Restraint: Favouring responsible growth over blanket bans.
- Fairness and Equity: Actively mitigating the linguistic and socio-economic biases of Western datasets.
- Accountability: Establishing clear liability for AI-driven outcomes.
- Understandable by Design: Mandating transparency so AI decisions are not “black boxes."
- Safety and Sustainability: Balancing high-speed processing with environmental stewardship.
Together, these principles seek to create guardrails without choking the momentum of emerging AI ecosystems, especially in developing economies.
Ending ‘AI Extractivism’
A defining feature of the declaration is its strong stance against “AI Extractivism” — the pattern in which data from developing nations is harvested to train powerful AI models, only for those same nations to purchase access to the finished products.
The Delhi Declaration aims to reverse this imbalance. By linking AI systems with robust Digital Public Infrastructure (DPI), it reinforces the idea of “Data Sovereignty,” ensuring countries retain control and economic value from their own data. The goal is to build domestic “Sovereign AI” capabilities rather than deepen technological dependence.
The Three Sutras: People, Planet, Progress
To translate principle into practice, the framework organises global collaboration around three measurable priorities.
People: The declaration promotes “Population-Scale” AI solutions. One example is the BharatGen model, which supports 22 Indian languages, addressing the reality that the vast majority of the world does not operate in English.
Planet: Under the banner of “Green AI,” the pact encourages energy-efficient computing infrastructure and shared climate-modelling data to help vulnerable nations respond to environmental challenges.
Progress: To democratise access, the declaration proposes a global “Compute Bank,” inspired by India’s initiative offering high-end GPUs at subsidised rates of Rs 65 per hour — a move designed to lower entry barriers for startups and researchers.
By blending development priorities with governance safeguards, the Delhi Declaration positions February 19 not merely as a summit date, but as a milestone in reshaping the global AI order.
