Relevance: GS-3 (Science & Technology – AI Regulation), GS-2 (Governance)
Source: The Hindu; DPIIT Working Paper

A new DPIIT working paper proposes that Large Language Models (LLMs) should be allowed free access to all publicly available online content for training. Instead of restricting AI data scraping, the government suggests creating a copyright royalty society to fairly compensate content creators. The proposal opens a major debate on how India should regulate Artificial Intelligence.

What Are LLMs?

Large Language Models are advanced AI systems trained on massive text datasets to recognise patterns and generate human-like language.
Examples: ChatGPT, Gemini, Claude.
They rely on web-scraped content, raising concerns about:

• Copyright,
  
• Consent,
  
• Data ownership,
  
• Ethical reuse of creative work.
  

What the Working Paper Proposes

• Free access for LLMs to publicly available content.
  
• A royalty-based model, not opt-outs, to avoid operational complexity.
  
• A government-recognised copyright society to collect and distribute fees to writers, publishers and media houses.
  
• Royalty distribution based on web-traffic, social indicators, and extent of content use.
  

Value Addition: This resembles the compulsory licensing framework used in Indian broadcasting and music industries, where users can access copyrighted material by paying a statutory fee.

Why This Matters for India

• Provides regulatory clarity for India’s AI ecosystem, reducing litigation.
  
• Supports AI innovation, an important pillar of the National Digital Strategy and Digital India Vision.
  
• Addresses publisher concerns—many media houses are already involved in copyright disputes with AI firms.
  
• Helps India position itself in global debates on AI fairness, transparency and responsible use.
  

Key Challenges

• Balancing innovation vs. creator rights.
  
• Ensuring small creators are fairly compensated.
  
• Preventing monopoly power among big AI firms.
  
• Addressing ethical issues such as deepfakes, plagiarism, and data misuse.
  

India seeks a middle path that supports AI growth while ensuring creators are not left behind.

Q. “Evaluate India’s proposed copyright framework for Large Language Models. How can it balance digital innovation with protection of intellectual property?”