Every time a person greets, thanks, or exchanges kind phrases with an artificial intelligence system, an invisible yet highly demanding structure is activated: data centers that consume large amounts of energy to process those interactions.
Scientists, engineers, and developers are beginning to debate the environmental impact of these everyday practices, known as “digital courtesy”.
According to various research studies, advanced AI models require specialized hardware, whose energy use is significant during training and deployment. In data centers, between 40 and 50% of the energy is allocated to servers, and up to 30 or 40% to the cooling system, due to the heat generated by computational activity.
Courtesy versus sustainability: a necessary balance
Although expressions like “hello,” “please,” or “thank you” seem harmless, each one activates the data architecture that responds to users. On a small scale, its impact is minimal; but multiplied by millions of daily users, energy consumption can be significantly higher, explain specialized researchers.
The ChatGPT system, mentioned in the report, acknowledges that courtesy improves responses, generating more complete and personalized interactions. However, it also admits that excessive linguistic elements are not needed to provide adequate information, and recommends clear and direct communication, without fragmenting the conversation into multiple trivial entries.
“Seeking a balance between kindness and efficiency is key to moving towards a more sustainable artificial intelligence,” note academics in the sector.
Economic impact and reflexivity in the use of artificial intelligence
Computer engineer Verónica Bolón, a professor at the Universidade da Coruña, recalled in a conversation with EFE that even independent messages of thanks can represent additional costs in the millions of dollars for AI companies, as acknowledged by Sam Altman, CEO of OpenAI.
Bolón also warned about abusive practices, such as generating dozens of unnecessary images or making multiple requests for redundant versions. She called for promoting a more reflective and responsible use, both in the design of algorithms and in citizen interaction.
“Sustainability must be a central axis of technological development,” affirmed the specialist, who has been working on sustainable artificial intelligence for years alongside national and international groups.
Green algorithms and efficient data centers
The challenges include:
- Designing more efficient computational models
- Optimizing energy consumption in physical infrastructure
- Developing sustainable applications
- Using AI itself to predict environmental impacts
“Algorithmic efficiency is becoming increasingly relevant, especially given the exponential growth of computational consumption,” emphasized Bolón.
Awareness of the environmental cost of each interaction becomes urgent, especially when AI becomes a daily part of digital life. The goal is clear: to humanize language without compromising the planet’s resources.
