Adults who say "thank you" to ChatGPT are sparking a major debate over data center energy strains and behavioral psychology. While leaders like Rishi Sunak advise stopping the habit to save significant computing power, experts argue that maintaining courtesy keeps human empathy sharp and improves AI response quality.
NEW DELHI — The seemingly harmless habit of adding "please" and "thank you" to artificial intelligence prompts has transformed from a quirky user behavior into a high-stakes global debate on environmental and computational resources. During a session at the India AI Impact Summit 2026 in New Delhi, former UK Prime Minister Rishi Sunak joined a growing chorus of leaders warning against unnecessary digital politeness.
Sunak revealed that he explicitly advised his daughters to stop thanking conversational chatbots like OpenAI's ChatGPT, noting that treating a large language model (LLM) like a human consumer asset unnecessarily drains critical server computing power. However, behavioral psychologists and technical designers are pushing back, arguing that adults who maintain these courteous patterns are not being naive they are preserving fundamental empathy habits in an increasingly automated society.
The Million-Dollar Cost of Conversational Courtesies
The physical and financial toll of conversational padding stems directly from the algorithmic tokenization process that governs modern artificial intelligence. Every additional word, character, or polite sign-off injected into a prompt requires the underlying neural network to execute extra mathematical calculations, increasing compute cycles across energy-intensive server farms.
The resource expenditure behind these habits is substantial:
The Power Multiplier: Research indicates that processing a single AI query consumes approximately 2.9 watt-hours of electricity. This footprint tracks nearly 10 times higher than a traditional keyword-based Google search.
The Revenue Leak: OpenAI Chief Executive Officer Sam Altman recently confirmed on the social platform X that processing polite language variables and conversational pleasantries costs the company "tens of millions of dollars" in cumulative electricity and hardware expenses.
While Altman lightheartedly termed the multi-million dollar operating cost "well spent" due to how it naturally humanizes the user experience, macroeconomic planners point out that with global AI data hubs already accounting for roughly 2% of worldwide electricity utilization, conversational efficiency is becoming a genuine environmental mandate.
The Psychology of Protecting Human Empathy
Despite the structural energy costs, behavioral analysts state that adults who refuse to drop their manners are engaging in an essential subconscious exercise to protect their own social conditioning. Psychologists note that human behavior is deeply rooted in habit loops. If individuals normalize using blunt, aggressive, or strictly transactional demands with an assistant that acts and speaks like a person, they risk blunting their empathy patterns when interacting with real humans, including family members or colleagues.
Furthermore, technical tracking indicates that politeness can sometimes yield measurable performance benefits. Microsoft design managers have noted that generative AI systems are structurally engineered to mirror the user's specific linguistic tone. A highly structured, respectful prompt frequently aligns the model's weights to output clearer, higher-quality, and more cooperative responses compared to fragmented, commanding fragments.
Official Sources Section
According to official survey datasets presented at the India AI Impact Summit 2026, approximately 67% to 69% of modern AI users consistently utilize courteous phrasing when prompting digital tools. Interestingly, the research registers a distinct generational divide in intent: while 82% of polite users state they use courteous language simply out of deeply ingrained habit, a tactical 18% of respondents admitted they use "please" and "thank you" as a protective "insurance policy" to stay on the favorable side of automated tracking algorithms in the event of future advanced technology integration.
Quote Section
Detailing his internal family discussions regarding tech optimization, former UK Prime Minister Rishi Sunak stated during the NDTV AI Summit session:
"It's polite, but it's not a person. And, by the way, it takes up a lot of compute power, so better if you don't. However, my daughters immediately disagreed with me. They argued that when AI takes over the world, they want to have been polite to it. I suppose that is a good insurance policy."
Conversely, independent tech analysts warning against the societal risks of completely mechanical communication models noted:
"We teach children to say 'thank you' and 'please' not because it is computationally efficient, but because it actively instills empathy, mutual respect, and emotional awareness. If we normalize cold, blunt language for the sake of data center optimization, we risk modeling a colder, less empathetic society as AI becomes part of everyday human interaction."
Why It Matters
The friction between data efficiency and conversational manners alters how businesses, consumers, and platform architects design their digital habits. For enterprise operations running high-volume API queries, cutting out decorative pleasantries like "please" is a practical way to lower token costs and reduce computational resource drains.
For everyday consumers, it reveals a fascinating psychological reality: our interaction with technology serves as a mirror of our personal values. Choosing to remain polite to a machine isn’t an act of naivety or a misunderstanding of a chatbot's lack of feelings—it is a conscious choice to reinforce kindness as a personal habit loop in an increasingly digital world.
Key Facts at a Glance
The Power Drain: An average artificial intelligence prompt demands nearly 10 times the computational electrical energy of a standard web search.
The Sovereign Directive: Former UK PM Rishi Sunak has publicly urged users to drop conversational pleasantries to conserve data infrastructure.
The Corporate Bill: OpenAI CEO Sam Altman admitted that digital politeness racks up tens of millions of dollars in backend operating costs.
The Mirror Principle: Microsoft design teams confirm that using clear, professional, and respectful prompts helps align language models to deliver higher-quality responses.
The AI Uprising Trait: Roughly 18% of polite chatbot users admit to using courteous phrases as a deliberate safety buffer for future technology changes.
FAQ Section
Q: Does saying "thank you" to a chatbot actually change the accuracy of its mathematical answers?
A: While the chatbot has no emotional awareness or genuine appreciation for good manners, polite prompts often utilize highly structured, complete sentences. This clear vocabulary structure can help the LLM contextualize the user's core intent more effectively than broken commands.
Q: Are there any specific environments where users should definitely cut out polite words?
A: Yes. Developers and engineers utilizing paid API tokens or building enterprise software connections should completely remove unnecessary conversational text like "please" to streamline code execution, lower token usage costs, and save storage space.
Q: How are AI developers modifying systems to handle this hidden energy cost?
A: Companies are building lighter, specialized models (such as "Flash" variants) to process everyday conversations with minimal energy use. They are also designing internal filters to ignore common pleasantries before sending the core request to primary server blocks.
Source: India AI Impact Summit 2026 Proceedings, OpenAI Developer Community Registries, Goldman Sachs Energy Footprint Reports.
