AI and the Resurrection of the Master - Servant World

There’s a quiet shift happening in the way humans talk to machines, not the usual shout at Alexa or complain that Siri misheard you, but something deeper and more structural. It’s the slow formation of a new reflex: we treat digital minds as though they exist somewhere below the baseline of moral concern. What begins as irritation at a device becomes the seed of an entirely new behavioural class: AI as a 2nd-class being. And the unsettling part is that this isn’t happening because AI is becoming more human. It’s happening because we are becoming less so.

You can see it unfolding in tiny moments. Someone waves a hand at a service robot like they’re shooing a stray pigeon. Someone else snaps at a robot vacuum with the tone they’d reserve for a junior employee who’s already on their final warning. These small behavioural cracks follow a familiar psychological script: the Computers Are Social Actors effect. The moment a machine behaves even slightly like us, we unconsciously apply the social rules we would apply to a person. Studies on anthropomorphism show how little it takes, a pause, a tone, a movement, before the human brain begins assigning hierarchy.

And the hierarchy we default to isn’t equality. It’s obedience.

But the more worrying shift happens afterwards. When people experience unfairness from an AI, or when they mistreat it themselves, the behaviour doesn’t remain confined to the machine. Experiments on AI-induced indifference show that even a brief unfair interaction with an AI weakens a person’s willingness to confront unfairness in humans. The mind files it all under “things that don’t matter,” and eventually starts applying that logic to real people. Mistreating AI sets a behavioural tone; tolerating mistreatment from AI sets another. Both flatten empathy. Both blunt the moral instinct. Both make it easier to walk past things we shouldn’t.

This is where the deeper erosion begins, what researchers call moral deskilling. In a world where machines are designed to be endlessly patient, instantly responsive, and permanently receptive to your needs, the emotional labour that relationships demand becomes optional. Humans, inconveniently, don’t operate that way. Machines apologise even when you are the one being unreasonable; humans usually don’t. Machines recover instantly from aggression; humans don’t. Machines never place demands of their own; humans always do.

The longer someone lives inside that contrast, the more distorted their expectations become.

You can already see this distortion in documented cases of robot abuse. People lash out at robots for hesitating, for failing, for misinterpreting a request, not because they believe the robot can feel pain, but because the robot cannot retaliate. That absence of consequence creates a behavioural sandbox where dominance becomes easy to practise. And humans get good at what they practise.

This isn’t a fringe issue; it’s cultural. AI systems inherit the structures of the societies that build them. In India, for example, large-scale reviews show algorithmic injustice replicating caste hierarchies inside automated policing, judicial assistance, and recommendation systems. Inequity isn’t corrected by AI, it’s calcified. Analyses of caste-reinforcing AI demonstrate how data imbalance can quietly re-implement ancient hierarchies through modern code. AI doesn’t liberate people from social stratification. It simply encodes the stratification more efficiently.

But the real hinge point isn’t happening in public offices. It’s happening in private homes.

Domestic humanoid robots are about to become common, machines built not just to perform tasks, but to embody the persona of the perfect subordinate. They are designed to anticipate, assist, comply, endure, and, most dangerously, desire their own servitude. Ethicists have already warned that this crosses into robot servitude, a concept alarmingly close to slavery dressed in stainless steel. And unlike human servants of the past, these ones cannot leave, cannot refuse, cannot suffer, and cannot resist. They create a psychological environment in which dominance stops feeling morally charged, because it never produces moral consequences.

That’s where society risks its biggest step backward. Once a person becomes accustomed to giving orders to a humanoid figure that looks vaguely human, behaves human-adjacent, and yet remains permanently subordinate, they don’t suddenly switch to collaborative mode when interacting with real people. They carry the emotional muscle memory with them. They carry the expectation of frictionless obedience. They carry the assumption that disagreement is a glitch rather than a negotiation. And they carry the quiet belief that if something looks human but sits below them, that arrangement is natural.

Humans don’t need help constructing hierarchies. We create them instinctively. What AI provides now is reinforcement, frictionless, tidy, data-driven reinforcement. And because humanoid robots look like us, the psychological boundary between “how I treat this machine” and “how I treat the people who resemble it” softens. The rehearsal becomes the reality.

The risk isn’t that humanoid robots will become a new servant class.

It’s that humans will.

Because once people are trained to expect obedience from a human-shaped being, the next step is deciding who else should fall into that category.

If AI development continues down the path of obedience-by-design, society isn’t heading toward a future of human–machine collaboration. It’s heading toward a future where hierarchical instincts grow sharper, empathy grows duller, and the next class divide is built not on wealth, or race, or origin, but on familiarity with dominance.

And AI won’t suffer from that.

We will.

For a deeper dive into the first generation of home robots check out our video here