Context:
Indian technology firms, supported by the IndiaAI Mission, are developing cost-effective Large Language Models (LLMs) tailored to Indian languages and local contexts, marking an important step in India’s AI ecosystem.
Key Highlights:
- Domestic AI Innovation
- Sarvam AI, a Bengaluru-based startup, launched two LLMs trained on 35 billion and 105 billion parameters.
- These models aim to support Indian languages and context-specific applications.
- Government Support
- Under the IndiaAI Mission, the government is commissioning over 36,000 GPUs across domestic data centres.
- Sarvam AI received access to 4,096 GPUs from the government’s common compute cluster.
- The subsidy provided is estimated at nearly ₹100 crore.
- Technological Breakthrough
- The models use Mixture of Experts (MoE) architecture, which activates only a subset of parameters during processing.
- This makes AI systems faster and more resource-efficient.
- Emerging Ecosystem
- BharatGen, incubated at IIT Bombay, has developed a multilingual 17 billion parameter model for sectors such as education and healthcare.
Relevant Prelims Points:
- Large Language Models (LLMs)
- AI models trained on massive datasets to generate and understand human-like language.
- Examples globally: GPT, PaLM, LLaMA.
- Parameters in AI
- Internal variables learned during training that determine model predictions and responses.
- IndiaAI Mission
- Government initiative led by Ministry of Electronics and Information Technology (MeitY).
- Aims to build AI infrastructure, datasets, and research capabilities in India.
- Graphics Processing Units (GPUs)
- Specialized processors essential for training large AI models due to high parallel computing capability.
- Mixture of Experts (MoE)
- AI architecture where only a fraction of the network activates during inference, improving efficiency.
Relevant Mains Points:
- Importance for India’s Digital Economy
- Indigenous LLMs reduce dependence on foreign AI models.
- Support digital inclusion by enabling AI tools in Indian languages.
- Governance and Strategic Implications
- Domestic AI development ensures data sovereignty and technological autonomy.
- Helps build secure AI ecosystems aligned with Indian policy priorities.
- Challenges
- High cost of computing infrastructure and energy consumption.
- Limited availability of quality training datasets in Indian languages.
- Need for ethical AI frameworks and transparency.
- Opportunities
- AI applications in education, healthcare, governance and agriculture.
- Strengthening India’s startup ecosystem and global AI competitiveness.
- Way Forward
- Expand public digital infrastructure for AI training.
- Develop high-quality multilingual datasets.
- Encourage open-source AI models and collaboration.
- Strengthen AI regulation and ethical standards.
UPSC Relevance:
- GS Paper III: Artificial Intelligence, emerging technologies.
- GS Paper II: Governance and digital public infrastructure.