Even the US Government Says AI Requires Massive Amounts of Water
A recent report by the Government Accountability Office highlights a surprising truth: generative AI systems guzzle an astonishing amount of water. With 250 million daily queries, these systems use over 1.1 million gallons, prompting concerns about the lack of transparency from companies regarding resource usage.
The 47-page analysis indicates that cooling the data centres, consuming a substantial 100-1000 megawatts of power, constitutes roughly 40% of their overall energy consumption. This percentage may climb as global temperatures rise. The report also reveals significant variability in water usage across different locations, with geographical factors greatly influencing both water needs and associated carbon emissions.
To illustrate, Meta’s Llama 3.1 model has emitted 8,930 metric tons of carbon, while Google’s Gemma2 model produced 1,247.61 metric tons, and OpenAI’s GPT-3 generated 552 metric tons. Importantly, the report suggests that generative AI searches come with costs approximately ten times higher than standard keyword searches, reiterating the ongoing transparency issues that render these AI systems “black boxes” to even their creators.
Key Points
- Generative AI systems use over 1.1 million gallons of water for 250 million daily queries.
- Cooling data centres accounts for about 40% of the total energy usage in AI systems.
- Water consumption and carbon emissions vary significantly depending on geographical location.
- Meta’s Llama 3.1 model has the highest carbon emission at 8,930 metric tons among major AI models.
- Generative AI searches cost approximately 10 times more than regular keyword searches.
- Lack of transparency continues to be a critical concern in AI system operations.
Why should I read this?
If you’re at all interested in the environmental impact of the technology we rely on daily, this article spills the tea on how our AI tools are guzzling water and emitting substantial carbon footprints. With everything going on about climate change and the tech industry’s role in it, knowing these details is vital. Dive in to catch the juicy facts that could make a difference in how we approach AI development and sustainability.