In contemplative classrooms, in conversations over a home-cooked meal, in discussions bouncing around a corporate office – the term AI has weaved its way into many types of rooms. The ethics of Artificial Intelligence – AI feeding off uncompensated authors, AI in the workplace, and AI information bias – are all noble and crucial discussions to be had. But what of the AI and human relationship?
Or, more pointedly, what are the ethical considerations surrounding our conversations with AI?

Source: PCMag Australia
I remember sitting in the backseat of my mother’s car, my friends and I hovering over an illuminated phone screen. One friend had recently discovered that if you asked Siri to divide zero by zero, she would give you the mathematical answer with an extra side of wit. Her tangent about dividing zero cookies amongst zero friends left the car overflowing with high-pitched giggles.
Until that same friend called Siri a name – one that the voices of eleven-year-olds don’t get to speak in front of adults – and the laughter bubbled down to a simmer. There were no consequences, no serious reprimands from Siri herself, yet the air had grown uncomfortable.
Siri doesn’t go on her little rant about the Cookie Monster anymore. And children should be given the grace to learn why “bad words” are exactly that. Yet, the confidence to mistreat Artificial Intelligence persists. I wouldn’t be writing this essay if the culprits were all that youthful.
Why does the way we speak to AI matter? Artificial Intelligence is just that – Artificial. There are no tears shed, hearts smashed, blood boiled. There seems to be no consequences – the AI we engage with is not designed to reciprocate cruelty or enforce punishment. There are no laws for us to fear.
However, there are always consequences. Only a foolish or arrogant man, seeking a false concept of freedom or running from responsibility, would ever assume otherwise. Unfortunately, we are all this man at times. It’s particularly easy to be this man when the consequences aren’t traditional, or tangibly visible.
Despite the research being over 20 years old, psychologists Justin L. Barrett and Amanda Hankes Johnson’s ‘The Role of Control in Attributing Intentional Agency to Inanimate Objects’ is as relevant as ever. Barrett and Johnson define intentional agents as “those objects that we understand as acting because of internal states such as beliefs and desires”. Our cognitive ability to detect intentional agents allows us to recognise other humans, acknowledge their autonomy, and distinguish AI as non-human and non-autonomous.

Source: Shutterstock
However, the very objective of AI is to mimic the intelligence and social behaviours of autonomous beings. ChatGPT has become a sacred companion for contemporary students, acting as a free tutor at the best of times and a confidant in procrastination at others. AI has become a teacher, an assistant, a therapist and, in some cases, a friend. As academics Andrea L. Guzman and Seth C. Lewis emphasise, such AI technologies “step into a [communication] role that […] historically has been restricted to humans”.
AI may not be an intentional agent, but it replicates the adaptive communication style that has only ever been attributed to beings with agency. It is vital that we recognise this, moving forward.
This is where the consequence lies. It is in the psychology of the human individual interacting with AI. Interactions and relationships have never been a one-way street, they are inherently a mutually influential experience. How an individual chooses to behave within their social environment can both construct the reality of their peers and redesign their own psyche. It is the concept of ‘manners maketh man’ in a nutshell.
In this case, it all comes down to respect. I am not referring to kneeling, head down and devout, before a computer. I am referring to the innate respect that all intentional agents – specifically, human beings – are entitled to. Respect as a term for the acknowledgment of a human’s individuality and autonomy, and the appropriate behaviour that results from this recognition.
So, while degrading Alexa from the first day she is unpackaged might not rewire her to seek your praise or resent you, it will – consciously or not – reaffirm your own idea that she is undeserving of basic respect. Whether or not she is necessarily owed respect is not an essential discussion here, simply because she is designed to perform in a likeness to autonomous beings that are.
By practising disrespect towards programs that simulate and replicate human interrelationships, we allow and condition disrespect to be exercised in reality.
It is also worth drawing attention to the intimate and complex relationship that respect has with control. Retracing our steps back to Barett and Johnson’s study, these psychologists found that individuals tend to grant agency to inanimate objects they do not have apparent control over. The flip side of this, however, is that when individuals are granted control, they tend to attribute less agency to what it is they control. The case is consistent with AI. We are not so quick to attribute agency to the predictable Grammarly, but Siri – who used to occasionally throw in an unexpected joke or two – would more often be described autonomously.
On the surface, this seems innocent enough. Attributing agency to the entities we believe to exhibit our own behavioural patterns is not necessarily harmful. Yet, definitive hierarchical systems of control establish a categorical space between positions. This space sets the foundation for potential division, to view the other person as a distinct ‘other’.
When control is involved, it is difficult to see ourselves in the ‘other’ and to identify our behavioural patterns within them, as they are either more constrained or more liberated by their power. Disrespect, mistreatment and an ignorance of agency are the best of friends, and this void of ‘otherness’ is their prime habitat.
This is why the ‘otherness’, the control, we have with AI makes our treatment towards it even more paramount. To close the void, to attempt to resolve these larger issues that permeate our human communities, we need to start at a basic level.
In our relationship with AI, we replicate human communication. We are in control. According to societal standards, we do not need to be respectful to AI. And yet, if we collectively and consciously habitualise respectful interactions with the non-intentional agents of AI, we ingrain the same respect for our fellow people. By continuing to deliberately recognise the control we have over AI, we cultivate a healthier community.
The way we converse with AI matters for ourselves, for our peers, and for the very endurance of respect in our technologically advancing world.
