Is AI a Blessing or Curse for the Environment?

• Artificial intelligence has often been promoted as a promising tool to combat climate change.
• However, tech companies should take steps to prevent artificial intelligence infrastructure from contributing to environmental destruction.

Artificial intelligence (AI) is rapidly becoming inherent in modern society. It is not just limited to chatbot operations but has also rapidly integrated itself into search engines, social media, and technologies such as GPS navigation. But, as the use of AI has risen, concerns have grown about the environmental cost associated with AI operations. OpenAI CEO Sam Altman has noted that the artificial intelligence industry is moving closer to an energy crisis.

Why AI Is a Resource-Hungry Beast

Each new version of a large language model (LLM) has been developed with massive improvements to parameters, enabling superior flexibility and accuracy. For instance, OpenAI’s GPT in 2018 had 100 million parameters, its next version had 1.5 billion parameters, and the third version had 175 billion parameters. Google’s PaLM LLM had 540 billion parameters. Consequently, these LLMs require several thousand high-performance chips and perform their operations through massive data centers.

Such data centers produce tremendous amounts of heat, and the cooling infrastructure is powered by fossil fuel energy sources. AI data centers are estimated to produce 2 to 4% of the global greenhouse gas emissions. Data centers also consume approximately 1% of global electricity resources, which is expected to at least double in the coming decade. According to Columbia Climate School, AI models consume significantly more when deployed than in their training phase, owing to the high number of users. This is a significant problem when a request to ChatGPT can exhaust 100 times more energy than a similar request in Google’s search engine.

In addition to energy needs, AI systems require massive amounts of fresh water to cool the processors. A lawsuit by Iowa residents against OpenAI’s GPT-4 facility revealed that the cluster consumed 6% of the district’s fresh water supply. This comes when many parts of the world face a critical water shortage crisis. The industry is also expected to generate significant amounts of e-waste.

With such issues becoming increasingly apparent, lawmakers are also taking note. In early February, US Senators introduced the Artificial Intelligence Environmental Impacts Act of 2024, requiring national assessment bodies to create frameworks to check AI’s environmental impact.

Additionally, the use of AI can cause other environmental concerns. For example, numerous leading fossil fuel companies use AI tech to bolster production. Using AI in targeted advertising also drives up the consumption of goods and services beyond what may be required for a sustainable society. With such a vast gray area in AI use and resource scarcity hitting multiple areas of modern society, can AI be seen as a blessing?

See More: How the Cloud Drives Sustainability

Areas Where AI Can Make a Positive Environmental Change

While concerns about AI’s environmental impact continue to grow, the technology also has tremendous potential in fighting climate change and controlling other sources of pollution. According to the World Health Organization (WHO), around 4 billion people live in areas vulnerable to climate change, which makes it essential that every resource available is used to stave off the crisis.

For instance, AI can play a significant role in reforestation projects, especially with other technologies like drones. They can carry out optimized seed bombing operations, particularly in locations that aren’t easily accessible. AI can also use satellite images to track deforestation and natural carbon stores.

Further, AI can be used with satellite imagery to track the movement of icebergs and the melting of ice caps faster and more accurately. A similar methodology can help heavy industries track and reduce their carbon emissions by up to 30%. Governments are increasingly using AI to monitor methane emissions and air quality.

Google is working with non-profit organizations that want to use machine learning in their projects. The United Nations is also using artificial intelligence to predict weather patterns, reforestation projects, and waste management for vulnerable communities in Africa to mitigate the effects of climate change.

AI is increasingly crucial in the operations of renewable energy infrastructure and smart grids. AI equipment can run at optimum efficiency while analyzing data coming from sensors and meters in real-time. The emerging technology can also significantly aid water management and sustainable agriculture, minimizing waste and improving produce quality and quantity.

With climate change becoming increasingly severe, AI can be used to predict the volatility of weather more accurately. This is essential for tracking natural disasters like unseasonal rain, hail, and storms. AI can also significantly improve waste management in terms of optimized recycling operations and ocean cleanup projects.

See More: 5 Keys to Attaining Sustainable IT 

Sustainability Guidelines for the AI Industry

AI has massive potential to solve the global climate change crisis. However, its current impact on the environment should not be ignored. Tech companies should take the initiative to apply their resources to solving the problems associated with artificial intelligence. Here are some ways AI can be made more environment-friendly.

• Government action: Governments are a key piece of the puzzle, with regulatory bodies pushing to establish sustainability and carbon emission laws while financially incentivizing companies to use eco-friendly practices.
• Transparency: AI companies can inform developers and companies about how much energy their AI use requires. This can lead to studies about the different AI systems and associated solutions to minimize energy use.
• Cooling systems: Although air conditioning is widely prevalent, it is not sustainable in the long term. Investing in researching alternatives can be a significant step toward improving the industry’s efficiency.
• Renewable energy: Investments in building renewable energy infrastructure are another step that can cut greenhouse gas emissions and drive sustainability in the AI industry.
• Optimizing and choosing models: AI companies need to optimize existing AI models instead of building new models with ever-increasing parameters, which could increase resource efficiency. Also, using smaller models for simpler tasks instead of LLMs for everything can bolster energy efficiency.
• Hardware and algorithm use: AI companies need to optimize their hardware to operate at lower speeds and set algorithms that need not produce results that are more accurate than necessary. Also, identifying a specific algorithm that is efficient in a particular task can drive sustainable AI use cases.

See More: Start Small: A Case for Incremental Change in Sustainability

As artificial intelligence solutions improve handling environmental problems, their value in fighting climate change could shine through. Climate policy and AI policies coming together will play a major role in the coming months and years. Establishing ethical principles for AI use to avoid climate harm needs to be part of every organization’s value system to maximize the potential of this new but volatile technology.

What do you think of AI’s role in sustainability? Share your thoughts with us on FacebookOpens a new window , XOpens a new window , and LinkedInOpens a new window . We’d love to hear from you!