The arguments over Artificial Intelligence’s (AI) impact on the environment are long-standing, and whilst we are unable to give you a definitive answer as to whether AI is good or bad, we can look at the ways that it is both helping, and harming, the environment.
How is AI Harming the Environment?
Electricity Usage
A large amount of electricity is consumed in data centres – temperature controlled buildings that house computing infrastructure such as servers – which are used to train and run AI. Global electricity demand from data centres could double in the next two years1, due to AI servers requiring up to eight times more energy than traditional computing solutions. Using large amounts of electricity can put pressure on the electricity grid, and if using fossil fuels, can result in harmful pollutants being emitted into the atmosphere.
Water Usage
Although not front of mind when thinking about AI, water usage is something to consider. As mentioned above, data centres require cooling infrastructure, using water to do so, which puts pressure on sourcing freshwater. This may put strains on local water supplies, and asks us to weigh up the need for water for AI products, when over 4 billion people worldwide do not have access to clean, safe water.
E-Waste
Although not directly linked to AI chat tools, e-waste is something to consider when we think of the rise of AI due to the electronic devices that are being produced to house and utilise AI technologies. Many electronic appliances contain hazardous chemicals, such as mercury, which contaminate soil and water supplies when buried in landfill, as materials dissolve into a sludge that penetrates to the bottom of landfill and sinks into the ground. Unfortunately, recycling of these devices is often complicated as they are difficult to dismantle, making the process labour intensive. The solution to the e-waste problem is better e-waste management and secure recycling, something that would need to be worldwide, and enforced from top level down.
Accountability
Unfortunately, many companies put their financial gains ahead of environmental effects when it comes to AI, wanting to win over customers with bigger and better systems, ignoring the impacts that these have on the climate around us.
Indirect Emissions
The increasing demand of AI applications results in a need for high-performing computing hardware, with the manufacture and transport of these impacting the environment, meaning negative effects at all stages of the product lifecycle.
How is AI Helping the Environment?
Measure and Predict
The beauty of AI is that it can look for patterns in data and use historic knowledge to measure sequences and predict future outcomes. One way in which AI tools are doing this is for electricity generation, being able to identify inefficiencies to optimise energy efficiency, and enhance forecasting of renewable energy sources. What’s more, AI can help to identify ways to reduce energy consumption and reduce energy waste, both for energy providers and end customers.
Additionally, AI can easily monitor air and water quality, allowing us to analyse small changes in the environment, to better understand the impact of everyday life around us, and to track climate change.
Manage Agriculture
AI can help with management of agriculture and transport. Not only can it help with transport movement, which in turn can be made more efficient to use less resources and create less emissions, but it can also be used to track and monitor endangered species and inform conservation strategies.
Do We Need to Use AI?
Looking at the environmental impact of AI begs the question, do we need to use AI in absolutely everything we do? Whilst there is no doubt that AI can benefit parts of our lives, we should be mindful of our individual interactions with AI chats. While a single ChatGPT query uses 2.9Wh of electricity, Google search uses only 0.3Wh2, meaning ChatGPT uses nearly 10 times more energy than Google Search, meaning it’s more important than ever to ask ourselves, “Could we Google our query rather than inputting it into an AI chat like ChatGPT?”.
Outlook
AI is a powerful tool with many benefits to reduce human errors and advance improvements in various fields and areas of life, however, its use should always be guided by ethical considerations. One of the main issues with AI is the volume of energy it uses to operate. However, this can be easily resolved if systems use energy from renewable sources instead. Data centres that run on renewable energy can run faster to make complex processes with AI more efficient, all whilst helping rather than harming the environment. The appropriate use of AI requires collaboration from governments, businesses and people to minimise its environmental impacts.
1 International Energy Agency
