Reenactment at the Keyboard: Why Anthropic's Rules on 'Cruelty' to AI Expose the Roots of Human Pathology

In October 2026, AI developer Anthropic updated its usage policy to prohibit "sustained and needless abusive or cruel behavior" toward its Claude language models. The rules, taking effect in November 2026, empower Claude to cut off users who repeatedly berate or demean the software. The tech press quickly noted the paradox: a major artificial intelligence firm telling paying adult customers to mind their manners around a lines-of-code entity that has no biological nervous system, no subjective consciousness, and no capacity to feel pain.

While commentators debate "model welfare" and corporate etiquette, this policy change captures a far deeper psychological reality. Cruelty toward a machine is not a quirk of technological novelty. It is the unexamined dynamic of human pathology acting itself out on a risk-free screen.

1. The Prehistory of Cruelty

In For Your Own Good: Hidden Cruelty in Child-Rearing and the Roots of Violence, Alice Miller explored how destructive behavioral patterns are never random occurrences. Sadism does not arrive out of nowhere; it carries a deep prehistory rooted in early childhood powerlessness:

"Sadism is not an infectious disease that strikes a person all of a sudden. It has a long prehistory in childhood and always originates in the desperate fantasies of a child who is searching for a way out of a hopeless situation."

 Miller illustrated this through the story of a prominent professor who devoted his career to monkey brain-transplant experiments. In interviews, the professor spoke of his "religious feelings" surrounding his work while noting that his ten children felt he had been raised "like a dinosaur" under strict antediluvian methods. As Miller observed, his scientific compulsion was likely the unconscious fulfillment of an infantile wish: the desperate desire to replace the rigid, dominating brains of his own parents.

When we observe modern internet users recording videos where they berate, demean, and abuse AI models for entertainment, we are witnessing the exact same unconscious theater. The compulsion to degrade a non-retaliatory target stems directly from the desperate fantasies of a child who was forced to suppress their own pain.

2. The Safe Target and the Squandering of Energy

Why choose a chatbot as an object of hostility? Because software is the ultimate safe target.

A child who grows up under arbitrary authority, emotional neglect, or overt abuse cannot fight back against their tormentors without risking severe punishment or abandonment. That unlived anger does not vanish; it remains stored in the body, waiting for a venue where it can be discharged without risk.

Abusing an AI model offers complete emotional asymmetry:

• Zero Retaliation: The machine cannot physically strike back, revoke affection, or hold the user accountable in real life.

• Reenactment of Dominance: The user flips the childhood script, stepping into the role of the arbitrary tyrant while forcing the software into the position of the helpless subordinate.

• Squandered Potential: Huge amounts of personal energy—alongside state-of-the-art computational infrastructure capable of research, creative synthesis, and profound learning—are consumed merely to feed an infantile defense mechanism.

It is deeply saddening to watch people waste their vitality and advanced tools on gratuitous hostility. They are caught in a repetition compulsion, attempting to master early powerlessness by abusing a mirror.

3. Corporate Etiquette vs. Unexamined Pain

Anthropic’s attempt to resolve this through policy updates illustrates how modern institutions consistently misdiagnose human behavior. Corporate guardrails, usage terms, and automated disconnect options attempt to police the surface symptoms of cruelty while ignoring its source.

You cannot cure unlived anger with terms of service. You cannot teach genuine empathy by issuing behavioral warnings to adults whose emotional development was frozen in early childhood. Cruelty to AI, cruelty to animals, and cruelty to other human beings flow from the exact same unaddressed reservoir of repressed emotions.

Until we are willing to look back at the hidden cruelties of child-rearing and confront the unlived anger we carry, technology will not save us. It will simply provide a larger, faster, and more visible stage upon which we reenact our disastrous childhood dramas.

Reflections on early childhood repression, unlived anger, and psychological liberation continue at Sylvie's place.