Generative artificial intelligence tools are now widely used, but recent studies indicate they have substantial hidden environmental costs. Testing 14 large language models (LLMs), researchers found that complex questions led to six times more carbon dioxide emissions than simpler ones, with advanced models emitting up to 50 times more. The energy-intensive processes driven by larger models' numerous parameters demands significantly more power, reflecting a trade-off between accuracy and environmental impact. Understanding this energy consumption is crucial as reliance on AI increases in various sectors.
Complex questions posed to AI systems can produce up to six times more carbon dioxide emissions than concise prompts, highlighting an environmental cost to accuracy.
Smarter, larger language models can generate up to 50 times more carbon emissions than simpler models, showing the trade-off between energy use and performance.
Collection
[
|
...
]