In an age where technology has been developing at unprecedented speeds, the internet and smart assistants such as ChatGPT have become a normal part of our everyday lives that we barely even think about. While something as simple as typing a question into Google may not seem like much, AI can be surprisingly damaging to the environment. This article details these impacts.
AI, which is short for artificial intelligence, is an umbrella term that refers to any kind of technology that can process information. As its name suggests, AI is characterised by its ability to mimic human intelligence.
It is important to first establish that AI has many positive uses for the environment: whether it’s analysing weather data and historical climate patterns, predicting future climate trends, monitoring data collection or even predicting natural disasters, AI is already helping to save thousands of lives and minimise the economic losses from climate change-related disasters. On top of that, AI could be the key to helping businesses and governments make more environmentally conscious decisions and help the energy sector transition away from fossil fuels. With the incredible versatility and ever-increasing capabilities of AI, the possibilities are virtually endless.
With all of the amazing things AI can do and the many more things it could potentially do in the future, what could possibly be so environmentally harmful about it? The impacts fall under 4 main areas:
Manufacturing - the hardware needed for AI to function, particularly microchips, require huge amounts of natural resources and rare earth metals to build, which are often mined unsustainably; in fact, 800 kg of raw materials could produce as little as 2 kg worth of components. The process of manufacturing these components also requires huge amounts of water, which is already scarce in many parts of the world.
Maintenance - due to the huge amounts of energy consumed by AI computer systems, the machinery requires huge amounts of water to cool it down and prevent overheating. A recent study by the University of Massachusetts also found that the process of training popular AI models can produce around 280 tonnes of carbon dioxide, which is nearly 5 times the amount of CO2 a single car produces throughout its entire lifetime.
Energy sources - the computer systems that power AI consume huge amounts of electricity, the majority of which comes from burning fossil fuels; this is a major source of air pollution and carbon dioxide emissions which contribute to global warming.
Disposal - when a component is no longer usable, it is often improperly disposed of and ends up in a landfill. Many electronic components contain harmful chemicals that will start to leak as they break down, causing soil and water pollution.
As technology continues to advance, we need to regulate our use of AI for the sake of the environment. This could include making AI as energy-efficient as possible, companies having full transparency on their technology’s environmental impact as well as the implementation of AI-regulation policies. To quote the Scientific American magazine:
“David Rolnick, a computer scientist at McGill University, likens AI to a hammer: “The primary impact of a hammer is what is being hammered,” he says, “not what is in the hammer.” Just as the tool can smash things to bits or pound in nails to build a house, artificial intelligence can hurt or help the environment.”
If implemented in the right ways, AI can become one of our most powerful tools in helping to monitor and mitigate the effects of the climate crisis. At the same time, we need to be conscious of how using such tools could also damage the very environment we are trying to protect. Finding the perfect balance between technology and nature will therefore be crucial in our ongoing fight against climate change.
Credits:
Image - Namirial Focus
Information - UNEP, University of Yale, Scientific American, University of Illinois Springfield
Comments