Since its launch in late 2022, ChatGPT has exploded in popularity, becoming a common resource for many. In my Entrepreneurship class this year, we use ChatGPT to ask and answer crucial questions for our business models. Artificial intelligence (AI) models have become widespread, with the Saint Francis English department even experimenting with TrueMark, a classroom-focused AI model, this semester. Regardless of the ethics of artificial intelligence (on which I would require a whole other article to rant) there exists one especially looming problem with AI: its climate impact.
While not all AI contributes equally to environmental damage, two of the biggest factors are large language models (LLMs) and generative AI. Both require intensive training before the models are able to be used, which requires an enormous amount of energy and electricity. Studies have shown that each query for gen AI (which uses LLMs for more natural responses) takes around ten times as much energy as a query to a non-AI search engine, like Google Search.
The electricity and energy used for these queries go to servers that release greenhouse gases, which contribute to global warming. Many of the data centers (which house servers) that process searches are housed underwater in order to prevent the centers from overheating due to heavy traffic. As a result, the uptick in servers and centers uses water that could otherwise be used in communities in which water is becoming increasingly scarce.
In addition to the energy and water problem posed by AI, building the chips needed for AI causes electronic waste and relies on (sometimes unethical) mining of rare materials. The United Nations Environmental Programme stated in an article on the environmental impact of AI, “the microchips that power AI need rare earth elements, which are often mined in environmentally destructive ways, noted Navigating New Horizons. The second problem is that data centres produce electronic waste, which often contains hazardous substances, like mercury and lead.”
As more and more companies rely on AI, the environmental impact will only grow. Per management consulting company McKinsey last May, “In the latest McKinsey Global Survey on AI, 65% of respondents report that their organizations are regularly using gen AI, nearly double the percentage from our previous survey just ten months ago.” The increasing dependence on models that fuel greenhouse gases, electronic waste, and water scarcity will ultimately exacerbate the current climate crisis, leading to environmental devastation.
Coming out of the hottest year in recorded history in which global temperature exceeded the 1.5°C threshold above pre-industrial levels, we must turn to more environmentally sustainable technologies. Simple research or queries should be directed towards other search engines that require fractional amounts of energy. Especially in California, which both houses some of the largest tech giants and faces increasingly devastating annual fires as a result of climate change, it is time to find greener technology. Each year will be hotter than the last and more places will become inhabitable unless we address AI and other large contributors to the climate crisis.
