Google’s greenhouse gas emissions soared by almost 50% over the last five years, according to its annual environmental report. The tech giant has put the 48% increase down, in large part, to the electricity required to run data centres for its artificial intelligence push. Users may have noticed more AI integration into search this year, while the Alphabet Inc owned company is also developing its own generative chatbot - Gemini, the artist formerly known as Bard.
The high electricity demand is due to the Large Language Models at the heart of popular artificial intelligence tools. If you are wondering why energy requirements for LLMs are so taxing, it is because vast computational resources are needed in order to run them. As a result, current AI systems lead “to significant energy use and corresponding carbon emissions”, writes the Scale Down substack.