Ironically, the first result when you Google ‘impact of AI on environment’ is the AI overview stating it “can have a negative environmental impact” but “can also be used to help address climate change”. Is this rather meta response accurate - is AI a net positive for our environment? Or is it time to stop asking it to draft our essays and solve our problem sheets in the name of saving the climate?

On November 30th 2022, ChatGPT was released by OpenAI and quickly became the most widely used chatbot in Internet history, Now, it receives approximately 3.8 billion visits a month. AI is no longer just a robot for billionaires in Silicon Valley, an inaccessible concept for the majority, but ChatGPT as well as other chatbots have become increasingly common in our day-to-day lives. Used for a multitude of everyday tasks such as writing emails, many students find it essential for progressing efficiently with academic work – in fact, my Maths supervisor encourages it.

“AI is no longer just a robot for billionaires in Silicon Valley, an inaccessible concept for the majority”

However, this new reliance on technology begs several ethical questions, one about the value of human-completed work – is it disingenuous to use AI to write a birthday message? Why are we allowing AI novels and artwork to overshadow the creations of real people? Another ethical question facing us is about the environment. With the ever-present threat of global warming, we must evaluate the impact of AI and its implications for the future of the climate.

A large carbon footprint follows the development, maintenance and disposal of the infrastructure necessary for AI. In 2023, OpenAI researchers suggested that since 2012, the amount of computing power used for deep learning research has been doubling every 3.4 months, and in May 2021 a report commissioned by the European Parliament estimated that by 2030, information and communication technologies (ICT) will use 20% of the world’s total electricity. Generating an AI image uses about the same amount of energy as it takes to charge a phone. Alongside greenhouse gas emissions, the rate of consumption of clean water is also worrying. By 2027, the global AI demand is projected to account for 4.2-6.6 billion cubic meters of water withdrawal. This is more than six times the total annual water withdrawal of Denmark, or half of the United Kingdom. When 1 in 3 people do not have access to clean drinking water, this raises particular concerns about the morality of the overuse of AI.

“By 2027, the global AI demand is projected to account for 4.2-6.6 billion cubic metres of water withdrawal”

On an individual scale, cutting down on use of chatbots and turning off automatic AI features will help. However, this will help in the same way cutting down on red meat and cycling do, which pales in comparison to corporations burning masses of fossil fuels – an often-quoted statistic states that 70% of international carbon emissions come from just 100 companies. A pattern of countries dedicating large sums to AI projects is cropping up all over the world – at the AI summit in Paris in early February, French President Macron announced that France would be dedicating €109bn to invest in AI. This demonstrates how Europe is fighting to compete on a global scale with China and the US by keeping up with their technological advances.

Macron referenced Trump at the summit, saying “Here there is no need to ‘drill baby, drill’, it is just ‘plug baby, plug’,” suggesting that there is some aim to focus on more sustainable progress. Most countries have adopted some recommendations on the ethical uses of AI and the EU has adopted legislation to try to curb the environmental impact. However, the competition is ever-growing and it is likely we will see an equivalent of the ‘Space Race’ of the 20th century (when Kennedy told the chief of NASA he was not interested in space, but in beating the Soviets), where nations prioritise advancing faster than opponents over the preservation of the environment.

As quoted by AI itself, it supposedly has the potential to have a positive impact. This suggestion takes the form of models which can predict climate change. It is also much more efficient at spotting patterns and anomalies and using historical knowledge to make these predictions. Despite this, it is clear that unchecked mass developments in AI infrastructure are going to be detrimental to the environment, particularly through carbon emissions and use of water.

So, in the same way you don’t need to flush your loo once a week, you can continue to use ChatGPT to summarise your reading… but maybe stop asking it to generate pictures of ‘happy Cambridge student’ in your Instagram DMs.

Want to share your thoughts on this article? Send us a letter to letters@varsity.co.uk or by using this form.