The notion that omitting “please” and “thank you” in queries to AI platforms such as ChatGPT could contribute to environmental conservation is gaining traction. Proponents argue that minimising the length of prompts reduces energy consumption. However, experts suggest that while the idea appears logical, the reality is far more complex. The energy footprint of artificial intelligence is shaped more by the frequency and intensity of use than by the politeness of user interactions.
The Energy Cost of AI Queries
The mechanics of AI operation fundamentally differ from traditional digital services. When a user opens a document or streams a video, the bulk of the energy expenditure has already occurred during the data storage phase. In contrast, each interaction with an AI model requires a fresh computation, incurring a distinct energy cost each time. This “inference” process demands significant computational power, which is why AI systems are often likened to infrastructure rather than mere software applications.
As AI technologies proliferate, their energy demands are becoming increasingly pronounced. Research published in the journal Science reveals that data centres, which are the backbone of AI operations, represent a substantial portion of global electricity consumption. This trend is projected to worsen, with the International Energy Agency warning that electricity demand from these centres could double within the decade if current growth patterns persist.
The Broader Environmental Footprint
The environmental implications of AI extend beyond energy consumption. Data centres also necessitate immense amounts of water for cooling and demand considerable land and infrastructural resources. Such needs have local consequences; for instance, New Zealand, with its high proportion of renewable energy, may attract data centre investments. However, this does not eliminate the strain on local power grids or guarantee that renewable resources will meet the new demand without adverse effects.
When viewed through a systemic lens, the introduction of AI technology exacerbates existing pressures on regions grappling with climate change, population growth, and resource competition. The interplay of energy, water, and land use is intricate and often overlooked in discussions surrounding AI’s environmental impact.
The Misplaced Focus on Politeness
While the suggestion to reduce word count in AI queries may be well-meaning, it distracts from the more pressing structural issues surrounding AI’s environmental footprint. The conversation needs to shift from trivial behavioural modifications to substantial considerations about how AI infrastructure is interwoven with energy planning, resource management, and land use priorities.
The perception that AI exists in a vacuum, free from environmental costs, is a misunderstanding that can be detrimental to climate adaptation efforts. Instead of merely focusing on prompt phrasing, stakeholders must address how the energy and water demands of AI technologies interact with existing systems and societal needs.
Why it Matters
Recognising the environmental impact of AI is crucial in fostering a grounded dialogue about its role in our society. The popularity of the “please” myth signals a growing awareness that AI is not a disembodied entity; it is tied to physical infrastructures that consume vast resources. Acknowledging this reality paves the way for meaningful discussions about integrating AI into our energy frameworks, ensuring that its benefits do not come at an unsustainable cost to our planet. As we navigate the complexities of climate adaptation, understanding AI’s footprint will be essential in shaping a sustainable future.
