I lied to my supervisor last week. I told her I had a doctor’s appointment at 9am so wouldn’t be in until 10. I didn’t have a doctor’s appointment. I just wanted a lie in because I’d been up until 2am watching Westworld.

If you haven’t seen it, it’s a dystopian series that came out in 2016 about a wild west theme park filled with AI ‘hosts’. Rich people pay to go there to indulge in whatever they want – predominantly riding horses, shooting people, and sexual assault – without fear of any consequence. The ‘hosts’ are programmed in a way that they can’t harm the real ‘guests’, and get reset at the end of each horrible day, forgetting everything. Until eventually, cracks in the system mean the AI hosts start to remember and retaliate. It’s a bleak look at human behaviour. And the whole time watching it I was wondering… ‘the wild west looks very dry. I bet the amount of energy it takes to generate a theme park full of AI people is KILLING their local environment.’

Okay, I wasn’t thinking that. I was thinking more about the human psyche’s need for violence. BUT, I figured it was a more interesting lead into the topic of AI without resorting to using AI to come up with one.

I find technological advancements scary and hopeful in equal measure. On the one hand, the idea that AI can help with medical advancements and do things like spot diseases substantially earlier is amazing. On the other, reading the International Monetary Fund (IMF) say AI could affect up to 40% of jobs and make financial inequality worse is frightening. Especially seeing as I watched a friend use ChatGPT to write an introduction for an essay that came out disturbingly good. Not as witty as if I’d have written it, obviously, but still good enough to make me sweat.

But there’s another big issue AI brings that has been lost in the excitement of things like Facebook posts showing what we’d look like as an action figure. It’s the impact AI has on the environment.

While we casually use this technology for things like creating the Toys R Us version of ourselves, tools like ChatGPT are burning through energy. Professor Gina Neff of Queen Mary University London told the BBC the ‘data centres used to power it consume more electricity in a year than 117 countries.’

Personally, I am not good with technology. I’ve just figured out how to put music on my Instagram stories. So, to me, AI is a totally mystical thing. The repercussions of using it never crossed my mind.

So, how is AI bad for the environment? The demand for generative AI services from big tech firms and people wanting computers that house AI is skyrocketing, meaning there’s been an increase of demand to millions of data centres around the world.

They're used for most large-scale AI deployments, including cloud service providers. The impact these centres have on the planet is massive because of the sheer amount of raw materials they rely on. Data centres are made up of thousands of racks of computer servers, which use massive amounts of energy and need large volumes of water to keep them cool.

Boss of the Scottish data centre firm DataVita, Danny Quinn, described the difference in energy use between a rack containing standard servers, and one containing AI processors: ‘A standard rack full of normal kit is about 4 kilowatts (kW) of power, which is equivalent to a family house. Whereas an AI kit rack would be about 20 times that, so about 80kW of power. And you could have hundreds, if not thousands, of these racks within a single data centre.’

The amount of water needed to run these centres is concerning when a quarter of humans already lack access to clean water and sanitation worldwide. And, although some tech companies have opted for alternative ways to cool servers (such as air-cooling), some experts and activists worry AI could worsen water supplies. One estimate has revealed that, globally, AI-related infrastructure may soon consume six times more water than Denmark - a country of 6 million.

On a more day-to-day basis, the International Energy Agency reported that a request made through ChatGPT consumes 10 times the electricity of a Google Search. The increase in energy usage to run these data centres and AI models massively affects greenhouse gas emissions and makes climate change worse.

If you need an example, Earth.org revealed a study conducted by researchers at the University of Massachusetts that looked to determine how much energy is used to train specific popular large AI models. The results showed training can produce roughly 626,000 pounds of carbon dioxide - the equivalent of roughly 300 round-trip flights between New York and San Francisco and nearly 5 times the lifetime emissions of the average car.

In my wildly uneducated opinion, I would save AI for medical development and things that will tangibly help make the planet and our time on it better. The rest are luxuries. Leave the threat to job security and what we look like as action figures out of it.

And DEFINITELY don’t use it to create a wild west theme park populated by AI bots that rich people do horrible things to. Although, if that did happen and the robots (quite rightly) rebelled and took over the world, I could probably use that as a good excuse to come into work a little late. If the bots haven’t already taken my job.