The Environmental Impact of Artificial Intelligence
In 1997, an IBM-developed AI software became the first program to beat a human chess champion. In 2004, NASA landed two rovers on Mars that navigated the planet without human intervention. And in 2011, Apple launched Siri, the first popular virtual assistant.
For years, artificial intelligence has been used to make strides in medicine, to predict market trends, and to write and organise code, but it’s only within the last few years that this technology has found its way into our everyday lives. Today, you can find AI functions in almost every digital application, from Canva to Google Search. With common uses spanning content drafting, idea generation, formatting designs, note taking or scheduling meetings, AI is relied upon by people both at work and within their personal lives.
And this technology only keeps advancing. In 2008, Tony Stark’s virtual assistant, Jarvis, seemed an impossible feat, but Google’s recently launched AI assistant, Gemini, is getting pretty close. On Monday, China’s DeepSeek AI overtook Chat GPT as the most powerful AI model, delivered at a fraction of the cost. Microsoft, an eager adopter of generative AI, plans to invest $80 billion in AI this year, while Meta announced at least $60 billion in investments on Friday.
But what is the environmental impact of AI and should we really be using it every day?
From water usage to energy consumption, AI has significant consequences for the planet.
Golestan Radwan, Chief Digital Officer of the United Nations Environment Programme, said: “There is still much we don’t know about the environmental impact of AI but some of the data we do have is concerning. We need to make sure the net effect of AI on the planet is positive before we deploy the technology at scale.”
The majority of largescale artificial intelligence deployments are housed in huge data centers that require enormous amounts of water to keep cool. According to research conducted by The University of California, Chat GPT uses a bottle of water per 100-words. In 2021, Google’s data centers alone consumed just under 13 billion litres of fresh water, at a time where droughts are common and severe.
In addition to this, an inordinate amount of electricity is required to power these data centers which is largely produced by fossil fuels. The average request made through Chat GPT, for example, consumes 10 times more electricity than a Google search. The International Energy Agency estimates that by next year, electricity consumption by data centers, cryptocurrency and AI could reach 4% of annual global energy usage.
But it’s not black and white. On the other side of the coin, AI can actually be used to help with climate change. It can detect patterns in data to predict future outcomes which is invaluable for monitoring the environment and helping governments, organisations and individuals reduce their environmental impacts. AI is already being used to chart greenhouse gas emissions, recycle more waste and predict weather patterns. In addition to this, AI can support process and supply chain optimisation to reduce environmental impacts.
The Sustainable Future of AI
Research surrounding AI’s climate implications is rapidly gaining momentum. While many of the industry’s big players have been reluctant to disclose environmental impact reports, growing social concerns will increase the pressure on these organisations to address and reduce the footprint of their AI models.
As the technology advances, efficiencies will also be made to energy consumption. China’s newly launched AI model, Deep Seek, relies on a lighter model than other leading AI products, which may offer a pathway to reducing AI emissions.
Regulators also have a say in advancing sustainable AI practice. The European Union’s recently approved EU AI act envisages the creation of voluntary industry standards for decreasing energy consumption for AI systems, as well as developing energy-efficient AI training methods. The Act will also establish that AI providers should disclose information on the energy consumption of their models.
Many governments are beginning to address the environmental impact of AI at a larger scale, but what can we do as users to reduce our environmental impact?
I think the key here is using it responsibly. Many of us take measures to reduce our personal climate footprints, from recycling to using reusable cups, so it’s not out of the question to suggest that we cut down our AI usage too.
While many of us rely on AI to assist with our roles, there are a lot of instances where using it isn’t essential. For example, generating unnecessary AI images, drafting social media posts, or using it as a search engine.
AI is a revolutionary technology that can help us all take huge strides in our work, improving productivity, aiding with strategic thinking, and streamlining processes. It’s not a question of cutting it out entirely, but of using it responsibly and efficiently.
Instead of using AI to generate quick images to accompany articles or social media posts, support freelance illustrators for work that’s unique to your business. Instead of relying on AI to draft posts for your LinkedIn account, write with your own words to generate content that’s true to your voice and doesn’t contain the tell-tale signs of AI content.
In terms of increasing efficiency, AI training is a great way to get the most out of this technology, from successful prompt-writing to knowing what tools to use and when.
Climate change is accelerating at a rapid pace and although the primary onus falls on AI development companies to run as sustainably as possible, and for governments to implement regulations and policies to protect against environmental damage, it’s important we as users are at least aware of the impact of the technology we’re using day-to-day.
