Tanya Weaver Fri 29 Nov 2024
Collected at: https://eandt.theiet.org/2024/11/29/how-boil-egg-and-other-simple-searches-chatgpt-worse-environment-you-may-think
Large Language Models (LLMs), such as ChatGPT, consume just over one 500ml bottle of water per 100-word request, research by the IET has found.
ChatGPT has turned two. On 30 November 2022, OpenAI released a statement on its website saying “we’ve trained a model called ChatGPT, which interacts in a conversational way”.
The chatbot quickly garnered attention for its detailed responses and articulate answers across many domains of knowledge.
Two years on and the IET has carried out research to uncover just how many people are using this AI tool.
The study found that almost half of the UK public surveyed (49%) use it casually, with two in five (39%) using it regularly at work.
The findings show that on average we use it once a week for tasks including writing emails (50%), creating presentations (48%) and project planning (48%) – as well as general home tasks such as writing shopping lists (46%).
Almost half (46%) believe the technology will make their lives easier, with 50% seeing themselves using it more in the future.
However, what toll does this use have on the environment, especially when asking it simple queries such as how to boil an egg or even the meaning of life?
Data centres play a crucial role in training and operating the models that underpin AI models such as ChatGPT. But to do this, they consume huge amounts of energy – a significant proportion of which is used to cool the enormous amount of waste heat generated during computation. They also require significant amounts of water, stored in cooling towers, to cool the servers and keep systems at optimal temperatures.
The IET research found that fewer than one in six people in the UK (16%) are aware of these environmental impacts.
To put it in perspective, the study found data from The Verge that revealed that training ChatGPT 3 is estimated to have used 1,300MWh of electricity, which is the equivalent of 1,625,000 hours of watching Netflix.
In terms of water consumption, analysis from the University of California found that ChatGPT consumes just over one 500ml bottle of water per 100-word request.
While that may not sound a lot, if each of London’s 9.7 million residents asked ChatGPT to write a 100-word email this would total 4,874,000 litres – equivalent to filling over seven 25m swimming pools.
Professor Peter Bannister, chair of the IET’s Sustainability and Net Zero Policy Centre, said: “We’ve seen a colossal uptake of AI tools in the past two years, and while LLMs can provide great support to users, there is significant cost and environmental impact, which is only going to increase as the adoption of AI rises to drive efficiency and productivity.”
The IET’s best practice tips for using LLMs more efficiently and effectively include:
- LLMs are most effectively used for producing text, so use them to assist with writing documents.
- Consider the searches you are inputting – can you find the information via a traditional search engine, which is far less intensive on the environment?
- If you need to use LLMs, think about the prompts you are inputting for your request – provide as much detail as possible into it so that you get as much out of it. The result will only be as good as the information you initially provide.
- Consider the environmental costs of LLM searches – it is reported that each ChatGPT query consumes an estimated 2.9 watt-hours of electricity, nearly 10 times more than a standard Google search.
As world leaders move to regulate the technology, the IET believes regulations should look beyond the immediate risks of AI development to the much broader impact it has on the environment.
Dr Graham Herries, chair of the IET’s Digital Futures Policy Centre, said: “This could include a bronze, silver or gold standard for the approval of new data centres – which support AI servers – in the UK, based on a sustainability rating.
“By focusing on carbon efficient AI and technologies, the UK has the potential to become a leader in the global AI economy.”
Leave a Reply