Researchers are advising users to make AI prompts more concise by removing polite filler words to reduce energy consumption [1].
This shift in user behavior comes as the global community faces mounting pressure to manage the environmental impact of large-scale computing. Because every token processed by a large language model requires computational power, reducing the volume of unnecessary text can lower the overall electricity demand of these systems [1].
Concise prompting means omitting social niceties, such as "please" or "thank you," when interacting with AI interfaces. While these phrases are common in human conversation, they serve no functional purpose in guiding an AI to a correct answer [2]. By stripping away this filler, users can decrease the number of tokens the model must process to generate a response [1].
The recommendation follows a warning from a UN report regarding the rapid growth of AI energy consumption [1]. The report indicates that the escalating power needs of artificial intelligence pose a significant challenge to sustainability goals [2].
Researchers said that improving efficiency at the user level is one way to mitigate these effects [1]. While individual changes may seem small, the cumulative impact of millions of users shortening their prompts could lead to a measurable decrease in the energy load on data centers [2].
This approach focuses on the input side of the AI equation. By streamlining the prompts, the computational overhead for each request is minimized, thereby reducing the carbon footprint associated with each single query [1].
“Removing polite filler words from AI queries may curb the rapid growth of energy consumption.”
The push for concise prompting reflects a growing tension between the rapid adoption of generative AI and global climate targets. As the UN warns of surging energy demands, the responsibility for efficiency is being shared between the developers who optimize models and the users who interact with them. This suggests that 'prompt engineering' is evolving from a tool for better accuracy into a strategy for environmental sustainability.
