INDIA’S FOUNDATIONAL AI MODEL INITIATIVE: SCIENCE & TECHNOLOGY

INDIA’S FOUNDATIONAL AI MODEL INITIATIVE: SCIENCE & TECHNOLOGY

NEWS: Indian govt receives 67 proposals for domestic AI foundational models under AI mission

 

WHAT’S IN THE NEWS?

India is working towards developing its own foundational AI model by the end of 2025 under the ₹10,000 crore IndiaAI Mission, aiming to reduce reliance on foreign AI technologies. However, challenges such as high computational costs, GPU shortages, and data localization issues pose significant hurdles to this ambitious goal.

 

Artificial Intelligence and India's Foundational AI Model Initiative

Introduction to Artificial Intelligence (AI)

  • AI refers to the ability of machines to mimic intelligent human behavior and perform complex tasks.
  • These tasks include thinking, perceiving, learning, problem-solving, and decision-making.
  • AI has applications in various sectors, including healthcare, finance, education, and defense.

 

Key Terms Related to Artificial Intelligence

1. Artificial Intelligence (AI)

  • A branch of computer science focused on creating machines capable of intelligent behavior.
  • AI-powered systems simulate cognitive functions such as learning, reasoning, and self-correction.

2. Machine Learning (ML)

  • A subfield of AI that allows computers to learn from data without explicit programming.
  • Uses statistical techniques and algorithms to improve performance over time.
  • Artificial Neural Networks (ANNs) are a subset of machine learning models.

3. Artificial Neural Networks (ANNs)

  • Inspired by the structure of the human brain with interconnected nodes (neurons).
  • ANNs can identify complex patterns in data and improve decision-making processes.
  • Example: AlphaFold (AI model predicting protein structures).

4. Artificial General Intelligence (AGI)

  • Represents a higher level of AI where machines can perform any intellectual task a human can.
  • AGI enables machines to think and act autonomously through self-learning capabilities.

5. Foundational AI

  • Large-scale AI models trained on vast datasets to support multiple AI applications.
  • These models serve as a base for various AI-powered applications, including generative AI.

6. Generative AI

  • AI models that can generate new and original content using machine learning algorithms.
  • Capable of creating images, text, code, audio, and video based on natural language prompts.
  • Examples:
  • DALL-E – Generates AI-powered images.
  • ChatGPT – Creates human-like text responses.

7. Large Language Models (LLMs)

  • A subset of Foundational AI trained on huge datasets with at least one billion parameters.
  • LLMs exhibit high proficiency in understanding and generating human-like responses.
  • Examples:
  • ChatGPT
  • Gemini
  • Perplexity
  • DeepSeek
  • Grok

8. Small Language Models (SLMs)

  • Compact AI models with fewer than one billion parameters (millions to a few billion).
  • More efficient, cost-effective, and easier to deploy for specific use cases.
  • Best suited for localized or domain-specific applications.

 

India’s Challenges in Developing Its Own Foundational AI Model

  1. Limited AI Infrastructure & Computational Power
    • Foundational AI models require massive computational resources, including high-end GPUs.
    • India lacks sufficient high-performance computing (HPC) infrastructure.
  1. Dependence on Graphics Processing Units (GPUs)
    • GPUs are essential for training deep learning models.
    • Current shortage of GPUs worldwide is impacting AI development efforts.
    • India plans to procure at least 10,000 GPUs under the AI Mission.
  1. High Cost of Model Training
    • Training foundational AI models requires:
      • Thousands of GPUs running in hyperscale data centers (size: ~1 million sq. ft.).
      • High electricity consumption (Example: GPT-3 consumed ~1,300 MWh of power).
  1. Need for Large Datasets
    • AI models require huge and diverse datasets for training.
    • Data must be culturally and linguistically relevant to India to avoid foreign biases.
  1. Risk of Foreign Dependence
    • Building AI on foreign models poses security risks such as:
      • Vulnerabilities in defense and national security applications.
      • Data privacy risks due to reliance on external AI frameworks.
    • India needs its own foundational model to ensure data sovereignty.

 

India and AI: Government Initiatives

  1. IndiaAI Mission (2024)
    • Launched with an investment of ₹10,000 crore to develop AI capabilities.
    • Aims to develop India’s own LLM within 10 months (by the end of 2025).
    • Received 67 proposals for developing India-specific AI models.
  1. Evaluation of AI Model Proposals
    • A high-level technical committee will review and select AI model proposals.
    • Intellectual property of the AI models will remain with the developing entity.
    • Government will have a perpetual license for public use of the AI model.
  1. Procurement of GPUs
    • Government selected 10 companies to supply 18,693 GPUs.
    • GPUs will be distributed across research institutions and startups to boost AI development.
    • Initial target: Procurement of 10,000 GPUs for AI computing.
  1. AI Compute Facility for Startups & Researchers
    • A common AI computing facility will be launched to support startups.
    • Researchers and innovators will get access to high-performance computing power.
    • Government will provide a 40% subsidy on the total price of AI computing services.

 

Conclusion

  • India aims to reduce its dependence on foreign AI models by developing homegrown AI technology.
  • The IndiaAI Mission is a major step toward achieving technological self-sufficiency in AI.
  • By addressing challenges related to computing power, data availability, and security risks, India can emerge as a global AI leader.

 

Source: https://indianstartupnews.com/news/indian-govt-receives-67-proposals-for-domestic-ai-foundational-models-under-ai-mission-8746497