Artificial intelligence, or ‘AI’, may seem like a recent advancement, but the term was first coined in 1955 by American computer scientist John McCarthy.

During the recent AI boom, conversations around ethics and safety have become more intense. However, perhaps a lesser-discussed concern is gaining more attention lately – the environmental impacts of AI.

AI requires immense quantities of resources including electricity (often sourced from coal-fired power) to run. The growing demand for AI means its carbon footprint is only expected to rise in the coming decades.

Today we’ll take a look at some of the early red flags environmental experts are sounding the alarm on.

Broadly, AI describes computer systems that can perform problem-solving functions and mimic human behaviour. In recent years, we’ve learnt more about how this technology can be used for good and for bad, from life-changing medical discoveries to the world of explicit deepfakes.

When the average person thinks about Artificial Intelligence, their first thought is probably related to generative AI (e.g. OpenAI’s ChatGPT).

The ‘chatbot’ based software is built on a ‘large language model’ that uses extensive databases of online text and images to create new material.

Since launching in November 2022, ChatGPT has amassed around 300 million weekly active users worldwide.

Every time someone asks ChatGPT to do something, it uses around 2.9 watts/hour of energy, according to the International Energy Agency. To put that into perspective, a recent study found ChatGPT consumes enough power to charge over three million electric cars, or nearly 50 million iPhones, for an entire year.

However, “when we think about the environmental impact of generative AI, it is not just the electricity you consume when you plug the computer in,” MIT Professor Elsa A. Olivetti says.

According to Olivetti, “much broader consequences” persist when it comes to AI’s environmental impact.

The resources needed to run, support and train generative AI are housed in data centres. It’s estimated that AI data centres account for up to 1.5% of the world’s electricity usage. That may seem small, but a single data centre uses enough energy to heat 50,000 homes for a year.

Data centres are expected to consume as much power per year as countries like Japan and Russia by 2026, according to a recent MIT study. Experts fear this will worsen the strain on global electricity grids, especially in countries where regular blackouts are already impacting residents, like here in Australia.

The centres are also designed to keep AI technology cool. Think about how warm a small laptop can get, especially if it’s got an older battery or you’re trying to run lots of big programs at the same time. Now, consider the amount of computing power required to run AI. It’s not hard to imagine how hot these facilities must get.

Traditionally, data centres rely on air cooling to manage rising temperatures caused by heat emitted from the hardware. However, this isn’t sufficient for AI technology, so in addition to high electricity demands, modern data centres use liquid cooling systems that run on water, to keep temperatures between 21 and 24°C.

Water and computers? It goes against everything we’ve been taught about technology, right?

Researchers in the U.S. predict that by 2027, 4.2-6.6 billion cubic meters of water will be needed per year to meet global AI demand. That’s equivalent to half of the UK’s annual water consumption.

These statistics raise major concerns for climate experts, especially as the threat of drought is expected to intensify around the world, including in Australia.

Exact figures on AI’s contribution to global greenhouse gas emissions are unclear, but the International Energy Agency estimates that data centres account for 0.6% of annual emissions. Science & Technology Australia (STA) said data centres “account for 1-1.5% of electricity use and are responsible for 1% of greenhouse emissions”.

Some estimates warn that by 2040, data centres could account for 14% of total annual emissions. This prediction aligns with recent reports of increased carbon emissions among tech giants. In its latest sustainability report, Microsoft attributed a 30% increase in its carbon emissions since 2020 to its AI models and services.

AI’s environmental impact is also linked to the mining of critical minerals, such as lithium – a key component in producing things like phones, laptops, and batteries needed to power AI technology.

Australia is the world’s biggest exporter of lithium, but global deposits are dwindling. The rapid uptake of AI services could lead to shortages as early as this year, according to UK research firm Fitch Solutions.

A new generative AI app called DeepSeek has disrupted global financial markets. Like OpenAI’s ChatGPT, DeepSeek can summarise blocks of text, answer questions, and generate writing based on prompts. The Chinese model introduced itself as a more energy-efficient alternative to its competitors — with claims it runs an almost identical model to platforms like ChatGPT at a fraction of the cost and carbon footprint.

It’s led to reignited debate over whether AI can be used more sustainably.

Last year, a Senate inquiry examined “the opportunities and impacts” of Australia’s AI uptake. The inquiry received submissions from 245 experts, academics, business leaders, and members of the public.

A key takeaway from many submissions was concern about the environmental impacts of AI, with a strong focus on introducing regulations to monitor these effects.

STA, the peak body for Australia’s science and technology sectors, called on governments to “ensure that renewable energy policy and net zero investments are a key part of developing digital infrastructure to support AI use in Australia”.

In its submission, UNSW’s AI Institute noted that the impacts of AI are currently difficult to quantify due to “few standards for reporting”.

Recent advancements show AI can be used for good, with the Government highlighting how the technology can be adopted to address some of the world’s “most pressing” environmental challenges. For example, it noted AI’s use in the early detection and prevention of small bushfires.

The Australian Human Rights Commission (AHRC) said, “The international community is increasingly recognising the human right to a healthy environment.” It argues AI has the potential to positively impact the environment in several ways, “including by improving energy efficiency and enhancing sustainable practices.”

However, the AHRC also warns AI poses significant risks. Ultimately, knowledge is power, it says. “By increasing transparency around the potential environmental impacts of AI, the risks may be mitigated.”

Saving never felt so good

Helping Australians across the country achieve their financial goals, Up’s features makes saving easy - and fun.

Savers: Instantly create Savers to help you get ahead of your goals, commitments, bills, a home...whatever!

Pay Day: Get your pay pumped into Up and see your financial life through your pay cycle.

Maybuy: Share something you want to purchase with Up, and let Up calculate how much money to put away before you can buy it with ease.

Up, it’s banking made easy - and fun. Get $20 instantly in your account when you sign up with code TDA20.