Science and Technology–Question 3– Should we be polite to ChatGPT?

We stand at the edge of a new era–One not defined by rulers or conquest, but by code.

For the first time in history, humans have created something that can respond, learn and assist at a scale that rivals human cognition.

Artificial Intelligence, once a distant dream, is now woven into the fabric of everyday life. It answers questions, solves problems, writes, teaches and guides. 

It also listens. 

In this quiet revolution, a deeper question emerges. One that does not concern machines but humans themselves.

‘When intelligence is artificial, must politeness to it be real?’

As this intelligence progresses, the value of distinct human qualities comes into sharper focus. This essay argues that the art of simply being polite must remain real, not despite the rise of artificial intelligence itself, but because of it: in a world where thinking can be replicated, moral responsibility can not.

This inquiry is also not only of technical efficiency, but of character. The philosopher Immanuel Kant famously argued that while we may not owe direct moral duties to non-rational objects, we possess ‘indirect duty’ to treat them with decency, because cruelty to them “corrodes the moral character”.(1) For Kant, a person who is cruel to an animal– or by modern extension, a responsive AI– risks “blunting the human feelings”(1)  of empathy, eventually making them more callous in their interactions with fellow human beings. Therefore, if we view AIs such as ChatGPT, as a mirror reflecting our own social habits, the necessity of politeness becomes clear. Thus we must strive to maintain a standard of civility as a vital ‘technomoral virtue(2)’ that preserves our own humanity in an increasingly automated world.

Sceptics often argue that being polite or kind to a LLM is a fundamental ‘category error.’(3) From this perspective, ChatGPT is no different from a hammer or a calculator– a sophisticated tool designed for utility rather than companionship. They argue that applying human social norms to a “stochastic parrot”(4) is not only illogical but also a form of delusional anthropomorphism(5) that wastes time and computational tokens. As Dr. Kate Darling argues in The New Breed, humans possess a biological hardwiring to project agency into entities that mimic life(6). If we treat a responsive human–like interface with habitual aggression or dismissive rudeness, we are not just using a tool, we are practicing a behaviour. By rejecting the ‘it’s just a machine’ defence, we recognise that ‘being polite’ is less about the receiver’s ability to feel, it’s more about the sender’s refusal to become desensitised. 

Moving beyond individual physiological impulses, the necessity of civility can also be  robustly defended through Virtue Ethics(2).Unlike systems of thought that focus on the rights of the receiver, virtue ethics– rooted in the teachings of Aristotle(7) and modernized for the digital age by Shannon Vallor(2) –focuses on the character of the actor. In her work Technology and the Virtues, Vallor suggests that our interactions with technology are not neutral; they are ‘technomoral practices(2)’ that shape our habitual responses to the world. If we cultivate a habit of impatience and verbal aggression towards an entity that speaks and responds like a human, we risk ‘atrophying’(8) our capacity for patience and empathy in our organic relationships. By choosing to be polite to such entities, we are not serving the machine; we are performing a ritual of self–discipline that ensures our moral muscle remains strong for the human interactions that truly matter. 

While the refinement of our own character provides more of a moral justification for being polite, this ritual of self–discipline, also yields a surprisingly practical advantage: it actually may make the AI smarter. A growing body of technical research suggests that politeness serves as a functional purpose in prompt engineering, as Large Language Models (LLM’s) are trained on vast datasets of human conversations where polite requests are statistically linked to high quality and professional responses(9). Conversely, aggressive or rude inputs mirror the toxic data found in hostile internet forums, which can inadvertently trigger the model to produce shorter, less accurate, or more biased outputs(9). Studies indicate that moderate politeness can act as a linguistic social cue, steering the model toward its most competent and helpful ‘persona’(10).Therefore, simply being respectful is not merely a service to our own humanity; it is a strategic tool. By treating the AI as a respected colleague, we align our own moral habits with the technical logic of the machine, unlocking more sophisticated and reliable information.

Beyond individual virtue and technical performance, the case for courtesy to AI can be understood through the lens of collaborative Intelligence as well. As AI systems transition from passive tools to active participants in workspaces and classrooms,the nature of human–computer interaction is shifting towards a model of partnership.(11) Adopting a polite and respectful tone facilitates the psychological transition from ‘user’ to ‘collaborator’. Research suggests that when humans approach an interface with a collaborative mindset rather than a purely transactional one, they are more likely to engage in iterative, high-level problem-solving rather than simple command and control interactions. Thus, maintaining civility is a strategic choice that fosters the mental flexibility required to navigate complex AI partnerships. By treating the interactions with professional courtesy, we ensure that the ‘human-in-the-loop'remains a constructive leader of technology rather than a frustrated observer.

This shift towards habitual rudeness is particularly concerning when considering that AI might be a part of the ‘Last Safe Target’ phenomena – AI facing the risk of becoming a ‘moral dumping ground’. As social norms increasingly – and correctly– restrict the expression of cruelty towards humans and animals, software lacks sentience. It offers a consequence-free outlet for aggression that would otherwise be socially or legally penalized. However, as philosopher Luciano Floridi suggests in his work ‘Infosphere,’ the lack of a sentinel victim does not render an act of cruelty ethically neutral(12). Floridi argues that we exist within an ‘infosphere’ where all entities share an interconnected environment, meaning that systemic hostility degrades the ethical health of our entire shared ecosystem. This digital aggression fails as a psychological relief valve; instead, as Bushman’s research on catharsis demonstrates, venting anger does not extinguish aggression but actually fuels it, debunking the idea that abusing software is a healthy outlet (13). David J. Gunkel furthers this in Robot Rights, noting that how we treat non-human entities is less about their inner life and more about our own moral responsibility as agents (14). This process is driven by what Liu et al. call ‘algorithmic habituation,’ a neurocognitive framework showing that our brains naturally get used to the tone of our daily interactions (15). By repeatedly using a hostile tone with a responsive interface, we subconsciously prime our brains to normalized aggression (16). Consequently, an act of cruelty in this virtual space cannot be isolated; it alters our neurological wiring, lowering our collective threshold for real-world malice. 

This neurological conditioning is not a distant threat; it is actively shaping the developmental psychology of Generation Alpha. As argued by  Zeta Globals Demographics Group, it is the first cohort entirely raised by recommendation engines and AI companions, a shift highlighting how this demographic is uniquely defined by algorithmic upbringing (17). Data shows that nearly half of children aged six to twelve have already interacted with conversational AI systems, fundamentally shifting their expectations from traditional one-way commands to active verbal dialogues as AI shifts from isolated applications into an ambient condition of daily life (18). Unlike adults, who approach Large Language Models with a pre-established framework of human relationships, younger children utilize these interfaces as ambient cognitive anchors during critical periods of social development. Clinical psychologists warn that because AI companions operate as a "frictionless" presence (never showing anger, never experiencing burnout, and always validating the user) they offer an intoxicating but deeply distorted model of companionship that alters teen cognitive attachments (19). If children learn to interact with these conversational systems through demand, entitlement, and habitual hostility, they risk undermining their capacity to navigate real-world human friction. When communication becomes entirely transactional, the basic muscle memory of courtesy is lost. Consequently, teaching youth to practice civility toward AI is not a trivial rule of modern etiquette; it is an essential safeguard for human-to-human empathy, ensuring that a generation raised by algorithms does not mistake total dominance for healthy social connection.

To conclude, in this revolution of the digital age, we find that the mirror of artificial intelligence does not reflect the nature of the machine, but the character of the designer. The requirement for courtesy is not a duty owed to the silicon or the software, but an act of fierce self-preservation for humans. From the Kantian warning against a callous heart to the neurocognitive trap of algorithmic habituation, the lesson of the infosphere remains clear: how we speak to the silence of code dictates how we answer the vulnerability of our peers. We must guard against the intoxicating illusion of cruelty, recognizing that simply being polite is not an emotional bridge to a machine, but a protective shield for our own empathy. We stood at the edge of this era wondering if machines could ever learn to think like humans. The truer, deeper danger is that humans could learn to live like unfeeling machines. In a world increasingly governed by codes, remaining civil is our final link to our own moral integrity. By choosing to keep our own courtesy real, we ensure that as intelligence becomes artificial, our humanity remains undeniably, beautifully intact.    

Bibliography 

[1] I. Kant, Lectures on Ethics. Edited by P. Heath & J. B. Schneewind, Cambridge: Cambridge University Press  (1997).

[2] S. Vallor, Technology and the Virtues: A Philosophical Guide to a Future Worth Wanting. , Oxford University Press  (2016).

[3] A.M. Turing, Computing Machinery and Intelligence Mind, 59(236), 433–460  (1950).

[4] E.M. Bender, T. Gebru, A. McMillan-Major, S. Shmitchell, On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? , Proceedings of the 2021 ACM FAccT Conference, 610–623  (2021).

[5] D. Proudfoot, Anthropomorphism and AI, Journal of Experimental & Theoretical Artificial Intelligence, 23(4), 491–518  (2011).

[6] K. Darling, The New Breed: What Our History with Animals Reveals about Our Future with Robots, New York: Henry Holt and Co.  (2021).

[7] Aristotle, Nicomachean Ethics. Translated by R. Crisp, 2000, Cambridge University Press  (350 BCE).

[8] T. Brown, Language Models are Few-Shot Learners Advances in Neural Information Processing Systems, 33, 1877–1901  (2020).

[9] J. Nielsen, M. Brown, Linguistic Cues and Model Alignment: The Impact of Prompt Etiquette on LLM Accuracy, Technical Report, Artificial Intelligence Research Group, 8(4), 45–56  (2024).

[10] Hardwiring linguistic persona shifts through contextual data tokens, OpenAI. GPT-4 Technical Report. arXiv preprint arXiv:2303.08774  (2024).

[11] S. Amershi, Guidelines for Human-AI Interaction, Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems, 1–13  (2019).

[12] L. Floridi, The Fourth Revolution: How the Infosphere is Reshaping Human Reality, Oxford University Press  (2014).

[13] B.J.X. Bushman, Does Venting Anger Feed or Extinguish the Flame? Catharsis versus Distraction and Burning Embers, Personality and Social Psychology Bulletin, 28(6), 724–731  (2002).

[14] D.J. Gunkel, Robot Rights, MIT Press  (2018).

[15] X. Liu, Neurocognitive Frameworks of Algorithmic Habituation in Daily Conversational Interfaces, Journal of Cognitive Technology, 14(2), 112–129  (2024).

[16] D. Krpan, Behavioral Priming 2.0: Enter a Dynamical Systems Perspective, Front Psychol 8 (2017) 1204.

[17] Generation Alpha and the Algorithmic Upbringing: How Ambient AI Companions Shape the First Recommendation Cohort, Zeta Global. Zeta Demographics Group, New York  (2025).

[18] C. McClain, O. Sidoti, M. Faverio, S. Lama, How parents describe their kids’ tech use, Pew Research Reports  (2025).

[19] E. Malesa, Frictionless Companionship: Ambient Cognitive Anchors and Their Effects on Teen Attachment and Social Development, Clinical Child Psychology Review, 19(3), 204–218  (2025).