The UK organisation for the business of Open Technology

“The Road to the AI Impact Summit: AI Openness Report” – Bangalore 2025

Executive Summary

This report focuses on the Road to the India AI Impact Summit and argues that AI’s future legitimacy, innovation and global access depend on embracing open source, open standards and open data, particularly for developing nations.

Openness drives global innovation and access

• Rather than trying to define “open source AI”, focus should be on the openness of the underlying components, software, models and data; the report introduces “Open Chakra AI”: AI supporting universal access in an open way across the Summit’s seven thematic Chakras
• The opening of Meta’s Llama 2 enabled China’s DeepSeek R1 to be built at 20% of the cost of previous LLMs, a leap of iterative innovation driven by openness; within days Hugging Face iterated again with R1-Open
• Prime Minister Modi’s Paris call for “access for all” through open source is critical to shifting AI from the hands of the few to the many and bringing the Global South into AI governance

Sovereignty, dependency and trust

• The report reframes sovereignty: the real concern is dependency on a few vendors, not the origin of technology, open source provides access that breaks dependency and enables interoperability and local control
• Approaches compared: the UK’s pragmatic, non-isolationist Compute Roadmap and global partnerships; India’s mission to build domestic capability with open AI technology and a Sovereign AI team; the EU’s Eurostack linking sovereignty to open architecture but risking fragmentation through “local source”
• Linux Foundation data: 79% of organisations regard sovereign AI as strategically relevant, open source is the leading approach (81%), and 94% see global cooperation as essential, sovereignty achieved through participation in open ecosystems

Data and application in healthcare

• Data remains the critical challenge for openness in AI; India’s datasets must reflect its linguistic, cultural and socioeconomic diversity, moving public data from PDFs to clean, structured, open datasets
• Case studies show open source platforms, Bahmni (deployed in 500+ hospitals across 50+ countries) and the Open Healthcare Network’s Care (200+ government hospitals across 10 Indian states, powering Kerala’s palliative care programme with 100,000 monthly home visits) enabling local adaptation, FHIR interoperability and data sovereignty

What the data tells us

• India is sixth globally in AI repositories with 1k+ stars (60, up 50% year on year, the fastest growth globally) and second in Gen AI contributors (695,000 in 2025); GitHub projects India will be its largest developer base by 2028
• The UK is fourth globally in AI repositories with 1k+ stars (149, up 32%), holds the world’s highest per-capita open source participation with over 4 million GitHub accounts, and Scotland-based AutoGPT is the number one AI repository globally
• The US is India’s and the UK’s top collaboration partner; the UK is India’s second-largest, a notable partnership given the UK’s size

A path forward

• Concrete Summit actions proposed: standardise the definition of openness (e.g. the Model Openness Framework), financially support maximally open Class I models, and develop open alternatives to proprietary chokepoints like NVIDIA’s CUDA