New study reveals massive carbon costs hidden in AI prompt usage
Researchers in Germany tested 14 large language model (LLM) AI systems and discovered how much energy and carbon emissions AI prompts consume. Complex queries produced up to six times more CO2 than short answers. Smarter LLMs with advanced reasoning generated 50 times more emissions than simpler models answering the same question.
Maximilian Dauner, the doctoral student leading the study at Hochschule München University of Applied Sciences, explained the tradeoff: 
“This shows us the tradeoff between energy consumption and the accuracy of model performance.”
These smarter models have tens of billions more parameters — the key processing units — than smaller models.
Dauner compared it to a neural network in the brain:
“You can think of it like a neural network in the brain. The more neuron connections, the more thinking you can do to answer a question.”
Longer answers and polite phrasing also increase energy use. Dauner urges users to be more direct and limit answer length to one or two sentences to save energy.
Sasha Luccioni, climate lead at AI company Hugging Face, added:
“Task-specific models are often much smaller and more efficient, and just as good at any context-specific task.”
Researchers warn many AI providers don’t disclose energy or server details, making precise carbon tracking impossible. Shaolei Ren at University of California, Riverside said:
“You can’t really say AI consumes this much energy or water on average — that’s just not meaningful. We need to look at each individual model and then (examine what it uses) for each task.”
Dauner suggests AI companies share carbon footprint per prompt to raise user awareness.
“Generally, if people were more informed about the average (environmental) cost of generating a response, people would maybe start thinking, ‘Is it really necessary to turn myself into an action figure just because I’m bored?’ Or ‘do I have to tell ChatGPT jokes because I have nothing to do?’”
Luccioni slammed the AI hype surge:
“We don’t need generative AI in web search. Nobody asked for AI chatbots in (messaging apps) or on social media. This race to stuff them into every single existing technology is truly infuriating, since it comes with real consequences to our planet.”
Transparency and efficiency remain key as AI usage soars globally.
