IAS/UPSC Coaching Institute  

14 Jan 2026 : India must focus on AI and its environmental impact

 Editorial 1: ​India must focus on AI and its environmental impact

Context

There are several sustainable AI practices that India can adopt.

 

Introduction

The use of Artificial Intelligence (AI) is widely discussed across sectors ranging from healthcare to agriculture. However, its environmental implications have received comparatively limited attention. According to an OECD working paper titled “Measuring the Environmental Impacts of Artificial Intelligence Compute and Applications”, the development of AI algorithms entails significant environmental costs, including a rising carbon footprint, which further intensifies climate change–related challenges.

  • The report highlights that the global ICT industry—including hardware such as televisions- accounts for an estimated 1.8%–2.8% of global greenhouse gas (GHG) emissions, while some assessments place the figure even higher at 2.1%–3.9%.
  • It also notes that data on the carbon footprint of AI models and their usage is often unreliable or incomplete, making accurate assessment difficult.
  • A Google report (August 2025) states that a single text-based AI prompt consumes only 0.24 watt-hours of electricity.
  • However, this seemingly low energy consumption has attracted criticism, with observers pointing out that the report’s findings are incomplete and potentially misleading.

 

Impact on the Environment

  • A September 2024 issue note by the United Nations Environment Programme (UNEP) warned that AI data centres could consume 4.2–6.6 billion cubic metres of water by 2027, increasing the risk of water scarcity.
  • The note also highlighted that training a single Large Language Model (LLM) can emit nearly 300,000 kg of carbon dioxide, reflecting the high carbon intensity of AI development.
  • A 2019 study titled “Energy and Policy Considerations for Deep Learning in NLP” found that training one large AI model releases over 626,000 pounds of CO₂, comparable to the lifetime emissions of five cars.
  • The operational use of AI further adds to emissions; a UNEP (July 2024) study reported that each query on ChatGPT consumes around 10 times more energy than a Google search, contributing to climate change pressures.

 

Solutions – The World

  • Global ethical recognition: In 2021, UNESCO adopted the Recommendation on the Ethics of Artificial Intelligence, highlighting the negative social and environmental impacts of AI; it was accepted by nearly 190 countries.
  • Regulatory initiatives: The United States and the European Union have emerged as leading jurisdictions proposing legislative frameworks to address the environmental footprint of AI, such as the Artificial Intelligence Environmental Impacts Act, 2024 and the EU’s harmonised AI rules.
  • Carbon-cost discourse: International debates increasingly focus on the carbon and energy costs of AI development and deployment, going beyond narratives that view AI only as a tool for environmental protection.

 

Solutions – India

  • Recognising AI’s environmental cost: India needs to formally acknowledge the environmental externalities of building and deploying large AI models, alongside discussions on AI for climate solutions.
  • Extending Environmental Impact Assessment (EIA): Since EIA is mandatory under the EIA Notification, 2006, its scope can be expanded to assess the environmental impacts of AI model development and deployment, similar to assessments for infrastructure or river projects.
  • Developing measurement standards: The government can create standardised metrics by engaging tech companies, think tanks, and NGOs to build consensus on terminology, indicators, and reporting requirementsfor AI-related environmental impacts.
  • Robust data collection: Systematic tracking of sustainability indicators—such as GHG emissions, energy use, water consumption, freshwater stress, and land-use impacts—should be undertaken to enable evidence-based policy-making on AI and the environment.

 

As Part of Disclosure Standards

  • The government may consider mandating disclosure of the environmental impacts of developing and deploying AI models under Environmental, Social and Governance (ESG) frameworks.
  • Such disclosures can be incorporated through regulations framed by the Ministry of Corporate Affairs and the Securities and Exchange Board of India.
  • This would ensure greater transparency and accountability regarding energy use, emissions, and resource consumption linked to AI systems.
  • India can draw lessons from the European Union, where the Corporate Sustainability Reporting Directive(CSRD) mandates disclosure of emissions from data centres and high-compute activities, including the training of large language models (LLMs).

 

Conclusion

The focus should increasingly move toward positioning AI as a contributor to global sustainability goals, while simultaneously addressing its environmental footprint. Adopting sustainable AI practices can help mitigate the adverse environmental impacts associated with AI development and deployment. These include greater use of pre-trained models to reduce computational intensity, powering data centres with renewable energy sources, and systematic reporting of AI-specific environmental metrics, such as energy consumption and emissions, to enable informed and responsible decision-making.