AI's Energy Appetite Soars, Could Rival A Small Country's Electricity Consumption
The adoption of AI tools in the times to come could lead to an increase in energy footprints. This could be intense enough to exceed the power demands of some countries
Artificial intelligence (AI) tools are being adopted at an unprecedented rate across industries. The adoption of AI technologies means more energy consumption and a larger footprint. According to a new study, AI's energy footprint might exceed that of a small country in the near-future.
What are the implications of AI tools?
According to a commentary published in the journal Joule on October 10, the adoption of AI tools in the times to come could lead to an increase in energy footprints. This could be intense enough to exceed the power demands of some countries, according to a Science Daily report.
Since the launch of OpenAI's ChatGPT last year, AI tools have taken the centre stage in the world of technology. Generative AI tools like ChatGPT and Bard are able to generate text, images, and more through simple text prompts. This way, redundant tasks may be eliminated.
How do AI tools consume power?
Such AI models heavily depend on large amounts of data that is fed to such models. While building on this, de Vries pointed out that Hugging Face, an AI company based in New York, has claimed that its multilingual text generating AI tool consumed about 433 megawatt-hours (MWH) during training. As Live Mint pointed out, it could power 40 average American homes for a year.
Also read: In The Era Of ChatGPT, This Indian Startup Is Using Artificial Intelligence To Help The Poor Earn Money
To generate content based on prompts, such AI tools use a considerable amount of computing power and energy. The Science Daily report stated that ChatGPT could cost 564 MWh of electricity daily to run.
"The result of making these tools more efficient and accessible can be that we just allow more applications of it and more people to use it," de Vries wrote.
Which countries consume as much power as AI?
In the years to come, AI servers will multiply at a mammoth scale. By 2027, AI-related electricity consumption could grow by 85-134 terawatt-hours annually based on projects. This is comparable to the annual electricity consumption of countries like Argentina, Sweden, and the Netherlands.
A study in July this year highlighted AI models' water footprint. According to that study, published by the University of Colorado Riverside and the University of Texas Arlington, ChatGPT gulps down 500 ml of water for every conversation.
Also read: AI Experts Issue Dire Warning, Say Artificial Intelligence Poses 'Risk Of Extinction'
As various tech companies attempt to outdo each other in the arena of AI tools, long-term consequences of such tools could intensify in the time to come.
What do you think about the intensification of generative AI tools? Let us know in the comments below. For more in the world of technology and science, keep reading Indiatimes.com.