- What is AI?
- Can AI help fight climate change?
- Does AI hurt the environment?
- How can AI be sustainable?
- Does AI help the planet or hurt it?
- What is the environmental impact of Google’s Gemini?
- The environmental impact of Meta AI
- What is the environmental impact of ChatGPT?
- What can you do to mitigate the environmental impact of AI?
- Summary
- AI users are expected to exceed 378 million
- Data companies have high hopes AI will solve a lot of the world’s environmental challenges
- Google, ChatGPT and Meta are just three forms of AI
Artificial intelligence (AI) has boomed in recent years, with loads of people now using it in one form or another every day, including simple Google searches, project planning and much more.
There could be as more than 378 million users worldwide, with some claiming the global active user base each day for generative AI is between 115 million and 180 million.
However, such big numbers beg the question: what is the effect of AI on the planet?
Below, we’ll be looking at various forms of AI and analysing its environmental impact and if there’s anything we can do to alleviate this.
What is AI?
Specifically, AI refers to technology that can process information and superficially mimic human thinking. According to the UN, some form of AI has been around since the 1950s, but in the past five years, it’s evolved at breakneck speed. This is largely due to technological advances and having data readily available, which is crucial for training AI.
The number of companies in the AI space is growing constantly. Here are some of the biggest names:
- IBM – IBM Watson
- Microsoft – Microsoft Copilot
- OpenAI – ChatGPT
- Google – Gemini AI
- Amazon – Alexa and Rufus
- Meta – Meta AI
- Adobe – Adobe Firefly
- Apple – Apple Intelligence
Can AI help fight climate change?
Yes, at least there are hopes that AI can help fight climate change. Right now, it’s being used to harness data to help optimise generation, distribution and consumption, which long-term, can help countries transition completely from fossil fuels to renewable energy.
It can also detect patterns in data and group together other prompts and questions others have used. For example, if someone asks AI on how to reduce your carbon footprint, it will scramble the internet and previous searches to find the right answer for you. It will then likely add your question into its database to help someone further.
Home heating is a huge contributor to climate change. If you’re interested in replacing your gas boiler, read our guide to the best heat pump and best air source heat pumps installers.
This pattern detection means AI can predict future outcomes, making it invaluable for monitoring the environment, renewable energy and the future of low-carbon technologies. This can then easily help individuals, businesses, governments and world leaders make stronger choices that will help each country meet their net-zero goals.
David Jensen, coordinator of the United Nations Environment Program’s Digital Transformation sub-program added to this and says that AI can help design more energy-efficient buildings, monitor deforestation and optimise renewable energy deployment.
“This can be on a large scale, such as satellite monitoring of global emissions, or a more granular scale, such as a smart house automatically turning off lights or heat after a certain time,” Jensen says.
Does AI hurt the environment?
Unfortunately, it is also the case that AI hurts the planet because it creates a lot of energy and uses a lot of water.
AI relies on huge amounts of information, all of which gets stored in data centres. It’s estimated that by 2026, data centres globally will consume 1,050 tWh of electricity. For context, this would be more than the whole of Russia and slightly less than Japan, making AI data centres the fifth biggest energy consumer in the world.
In 2022, data centres globally consumed 460tWh, putting them at 11th globally.
According to a report from Greenpeace, titled ‘Environmental Impacts of Artificial Intelligence’, data centres also require monumental amounts of water to cool down – about 2 litres for every kWh of energy.
Get free solar panel quotes
Answer a few quick questions, and our trusted installers will send you bespoke solar panel quotes – for free.
AI-specialised data centres alone will consume 338 billion litres by 2030, up from 30 billion in 2023. For context, water consumption for data centres globally will be 664 billion litres by 2030, meaning more than 50% will be taken up by AI. This is because AI data centres use water to cool by absorbing heat from computing equipment, and needs 2l of water to cool for every kW hour of energy used.
This is a problem because the world is facing a potentially catastrophic water shortage. The World Economic Forum said in 2023 that the demand for freshwater will by 40% greater than demand by 2030, the same year that AI’s water consumption is set to hit 338 billion litres.
The report also found that the average water usage rate (litres per kWh) will increase from 0.36 l/kWh (2023) to 0.48 l/kWh (2030) due to the higher power densities of AI data centres.
As well as water, AI also relies heavily on critical minerals such as copper, cobalt, nickel and rare earths, which are used to make semiconductors, batteries and microchips.
Get £7500 towards your heat pump installation through a trusted installer
An MCS-certified heat pump installer can apply for the Boiler Upgrade Scheme on your behalf, get a quote from one now
Copper mining on its own accounts for 0.2% of all global carbon emissions, and the International Copper Association expects this to increase, with demand for the commodity set to double by 2050, driven mainly by AI.
Sally Golestand Radwan, chief digital officer, UNEO said in September 2024 that while there is still much we don’t know about the environmental impact of AI, some data is “concerning”.
“We need to make sure the net effect of AI on the planet is positive before we deploy the technology at scale,” she said.
In short, AI could potentially hurt the environment because it relies heavily on:
- Energy
- Water
- Critical minerals
Greenpeace UK’s chief scientist Dr Doug Parr told us there remains “uncertainty” around AI’s affect on the planet.
“What is certain is that the boom in energy-hungry data centres is hoovering up huge amounts of water and energy at a time of growing strain on both resources,” Dr Parr said.
“Right now, huge data facilities are being waved through in the UK without a proper assessment of their impacts on local water systems and the power grid – that’s reckless by governments and regulators, and needs to change.
“The multi-billion-pound tech giants building new data centres should be forced to take some responsibility for funding solutions, whether it’s cooling methods using much less water or running on new, clean renewable power.
“It’s time for the government to lay down some rules and take responsibility instead of just cheerleading for the AI sector.”
How can AI be sustainable?
Based on all this, it might be easy to think that AI is simply bad for the planet. However, Greenpeace believes there is a way to make it sustainable and less harmful for the planet.
In its report which we’ve already gone into, the charity suggests a five-point plan to make sure the world can benefit from AI without harming the environment.
- Transparency and accountability – this requires honest and full information pn the energy and water usage and efficiency. Companies should also specifying how much energy and water is used by AI.
- Grid integration and adaptation to renewable energies – Greenpeace says companies must make sure their data centres do not put too much pressure on local power grids and guarantee a minimum of ‘waste heat’, that is heat that is generated but not used – should be reused for local heating networks.
- Regulatory adjustments – this is possibly the most important. According to Greenpeace, the EU’s AI law is entirely focused on human safety and ignores the possible effect on the environment. It says policymakers should change their laws to make sure companies investing in AI aren’t harming the environment.
Does AI help the planet or hurt it?
Back to our original question. The answer is a bit of both. There is no doubt that AI is a huge burden on the world’s resources, but it can also help businesses and politicians make more environmentally-friendly decisions and policies.
It is also possible that as AI evolves, everyday chatbots will become more efficient and be able to do everything we require of it by using less water and fewer microchips, with a substantially reduced impact on the environment than current models.
There are examples of this already. DeepSeek, a Chinese AI chatbot, caused chaos in US markets earlier this year when it was revealed it can do a lot of what ChatGPT does by using a fraction of the microchips – DeepSeek used 2,000 to build its model, the industry average is 16,000. This means DeepSeek is having a significantly smaller impact on the environment than others.
What is the environmental impact of Google’s Gemini?
Google’s Gemini is used across all of its devices, with a single text prompt using 0.24 watt-hours apiece, emitting 0.3g of CO2, consuming 0.24ml of water. Most recently, it launched a report to discuss its environmental impact and to dispel some myths.
In its report, Google said that as more people use AI, society will need to better understand how to soften the effect on the planet.
Google calculated its AI environmental footprint by comparing different models, including the hardware and energy it ran on, and measuring water usage, the effect of data centres overheating and much more.
Find out more here: AI tool is better for the planet than TV, Google claims
The environmental impact of Meta AI
Meta’s latest data centre was completed in 2024, according to the New York Times, and typically consumes around 500,000 gallons of water a day. The parent company of Facebook, Instagram and WhatsApp has 28 to 30 data centres currently operational or in development around the world.
The New York Times also reported that new data centres built to train more powerful AI systems are set to require around 1 million gallons of water per day.
Meta as a company emits around 7.5 million tonnes of CO2 equivalent annually, which is the equivalent of more than 1.4 million homes’ electricity use for a year.
There isn’t a lot of information or research around the environmental impact of Meta AI, but it’s at risk of increasing the company’s already high CO2 emissions.
The company is planning to build a new data centre, reportedly the size of Manhattan, dubbed Prometheus, and is planning to have it online in 2026.
What is the environmental impact of ChatGPT?
ChatGPT is probably the most well-known AI tool in the world. As a result, its environmental impact is big. This is because it requires energy and water to train its large language models. It then requires ongoing energy consumption to deal with any queries or text prompts.
Again, there’s little information on exact figures, but in addition to high energy consumption and data centre strains, the creation of ChatGPT also has an environmental cost, from the extraction of rare earth materials when it’s manufactured, similar to other AI models.
What’s more, according to UNEP, a request made through ChatGPT consumes ten times the electricity of a Google search.
What can you do to mitigate the environmental impact of AI?
It’s almost impossible to ignore AI, and you will almost certainly use it in some way, if you’re not doing so already. Our smartphones, laptops and tablets all use some form of AI. Even a simple Google search offers an ‘AI overview’.
If you don’t want to harm the environment, use it intentionally and sparingly. You can also support companies that use renewable energy for their data centres and manage your energy consumption to prevent ‘phantom energy’, which can help extend your device’s lifetime.
All these, can help mitigate the environmental impact of AI. Outside of this, even low-carbon technologies can go a long way in counteracting AI usage.
Summary
- AI refers to a group of technologies that can process information and superficially mimic human thinking
- It can detect patterns in data and group together other prompts and questions others have used
- From an environmental impact, there are high hopes that AI can help tackle some of the world’s biggest challenges
- Meta as a company emits around 7.5 million tonnes of CO2 equivalent annually, which is the equivalent of more than 1.4 million homes’ electricity use for a year
- Not using AI is virtually impossible in 2025. Our smartphones, laptops and tablets all use some form of AI. Even a simple Google search offers an ‘AI overview’
- Instead, use it intentionally and sparingly
Get free solar panel quotes
Answer a few quick questions, and our trusted installers will send you bespoke solar panel quotes – for free.
