Is AI Accelerating Climate Change? The Environmental Impact of Artificial Intelligence

Artificial intelligence (AI) has become a transformative force in technology, but its environmental footprint is raising concerns. Training and operating AI models like ChatGPT require massive amounts of electricity, leading to significant greenhouse gas emissions. This has sparked debates about whether AI is exacerbating climate change.

The Environmental Costs of AI

• Energy Consumption and Emissions: AI systems consume vast amounts of energy, contributing to fossil fuel emissions and pollutants. Since 2020, AI-driven activities have accounted for 0.6% of global greenhouse gas emissions.
• Water Usage: AI operations also demand substantial fresh water to cool data center servers. For instance, a single ChatGPT query can consume significantly more energy and water than a typical Google search.
• Growing Concerns: As wildfires ravaged California in early 2025, social media users highlighted AI's role in climate change, with claims that AI models like ChatGPT are "speeding up the climate crisis" and depleting water resources.

AI's Carbon Footprint

Research shows that training and using AI models generate considerable carbon emissions. For example, GPT-4, a model developed by OpenAI, produces roughly a quarter to half a pound of carbon emissions to write a 100-250 word email. Data centers, which power these AI systems, accounted for 0.9% of energy-related greenhouse gas emissions globally in 2020.

Water Consumption

AI's water footprint is another concern. A 2023 study estimated that GPT-3, the model behind ChatGPT, consumes about 500ml of water for every 10-50 responses. By 2027, global AI demand could withdraw 4.2-6.6 million cubic meters of water annually—equivalent to Denmark's total water usage.

AI's Potential to Mitigate Climate Change

Despite its environmental costs, AI also holds promise for combating climate change:

• Optimizing Energy Systems: AI can improve the efficiency of power grids and reduce energy consumption in buildings.
• Sustainable Infrastructure: AI aids in designing eco-friendly infrastructure and detecting methane emissions from oil and gas installations.
• Lower Emissions for Certain Tasks: Some studies suggest that AI-generated writing and illustrations produce fewer carbon emissions than human labor.

The Debate Continues

While AI's environmental impact is undeniable, its long-term effects remain uncertain. Scientists and tech companies are exploring ways to make AI more energy-efficient and sustainable. However, AI's use in industries like warfare could exacerbate environmental harm by accelerating resource destruction.

As Shaolei Ren, an AI researcher, notes, "We can measure the carbon emissions, but the benefit part depends on the usage scenario." The future of AI's role in climate change remains a topic of intense debate.