The Environmental Impact of Artificial Intelligence

Source-This post on The Environmental Impact of Artificial Intelligence has been created based on the article “Why AI’s present and future bring some serious environmental concerns” published in “The Indian Express” on 26 July 2024.

UPSC Syllabus-GS Paper-3- Awareness in the fields of IT, Space, Computers, Robotics and Conservation, Environmental Pollution and Degradation

Context– AI has the potential to transform climate change efforts. However, Google’s annual environment report revealed a 13% increase in its emissions footprint in 2023 compared to the previous year. This rise was mainly attributed to increased electricity consumption in data centers and supply chains due to the growing deployment and usage of AI tools.

What are the reasons for higher energy consumption of AI models?

1) Performing Complex Operation– They perform more complex operations than simple search engines, even when addressing the same query. They process vast amounts of data while formulating appropriate responses, resulting in more electrical signals for data processing, storage, and retrieval.

2) Heat Generation and Cooling Requirements -The increased computational work generates more heat, requiring stronger cooling systems in data centers and increasing energy use.

What are the Environmental Impacts of Artificial Intelligence?

1) Enhances Electricity Demand– Data centers currently use 1% to 1.3% of global electricity, and this could increase to 1.5% to 3% by 2026, according to the International Energy Agency. In some countries, data centers use over 10% of national electricity. In Ireland, it’s 18% due to tax incentives, while in the U.S., it ranges from 1.3% to 4.5%.

Read More- European Union adopts first AI Law

2) Energy Intensive– AI queries, like those to ChatGPT, use 10 to 33 times more energy than a regular Google search, with image-based searches being even more energy-intensive.

3) Water Resource Concerns- There are increasing concerns about water use for cooling data centers. For instance, a data center for OpenAI’s GPT-4 in Iowa used 6% of the local water supply in July 2022.

What should be the way forward?

1) Need for Sustainable Planning- As India adopts AI technology, experts stress the need for careful planning to reduce environmental impacts. This means data centers should use efficient processes and minimize emissions.

2) Benefits of AI in Emissions Reduction – Studies suggest that widespread AI use in corporate and industrial practices could cut global emissions by 5-10% by 2030, according to Boston Consulting Group.

3) Monitoring and Predicting– AI can cut emissions by monitoring, predicting, and optimizing processes to reduce waste and inefficiencies.

Question for practice

What are the reasons for higher energy consumption of AI models? What are the Environmental Impacts of Artificial Intelligence?