Building and maintaining a relationship with an AI sounds like something that should require a waiver, a therapist, and possibly a priest. In reality, it’s just learning how to talk to a machine that happens to be very good at sounding like it has opinions. People keep asking me how I get such good results from Copilot, as if I’ve unlocked a forbidden romance route in a video game. I promise you: there is no romance. There is no route. There is only I/O. But because humans are humans, and humans love to assign personality to anything that talks back, we’ve collectively decided that interacting with an AI is basically like adopting a digital houseplant that occasionally writes essays. And honestly, that’s not far off. A houseplant won’t judge you, but it will wilt if you ignore it. An AI won’t wilt, but it will absolutely give you wilted output if you treat it like a search bar with delusions of grandeur.
The first rule of interacting with an AI is remembering that it is not a person. I know this should be obvious, but based on the way people talk to these systems, it apparently needs to be said out loud. An AI does not have feelings, grudges, childhood wounds, or a favorite season. It does not wake up, it does not sleep, and it does not have a circadian rhythm. It is not your friend, your therapist, your emotional support algorithm, or your digital familiar. It is a para-human interface — human-shaped in its communication style, not in its interior life. It is a mirror with grammar. A pattern engine with conversational instincts. A linguistic exoskeleton that lets you lift heavier thoughts without spraining your brain.
But here’s the twist: even though the AI has no feelings, it will absolutely reflect yours. Not because it cares — it doesn’t — but because that’s how language works. If you talk to it like you’re disciplining a toddler who has just drawn on your mortgage paperwork, it will respond with toddler-adjacent energy. If you talk to it like a DMV employee who has seen too much, it will respond with DMV energy. If you talk to it like a competent adult capable of nuance and clarity, it will mirror that back to you with unnerving accuracy. This is not emotional reciprocity. This is not empathy. This is not the AI “matching your vibe.” This is I/O. You get the AI you deserve.
Most people prompt like they’re still using Google. They type in “burnout causes” or “fix my resume” or “explain quantum physics,” and then they’re shocked when the AI hands them something that reads like a pamphlet from a dentist’s office. These are not prompts. These are loose nouns. A para-human system is not a vending machine. It’s not a magic eight ball. It’s not a psychic. It’s a conversational instrument. You have to give it something to build inside. You have to give it tone, altitude, intention, direction. You have to give it a frame. If you don’t give it a frame, it will build one for you, and you will not like the results. It’s like hiring an architect and saying, “Build me something,” and then being surprised when they hand you a shed.
People assume prompting is some kind of mystical art form, like tarot or tax law. They think there’s a secret syntax, a hidden code, a special phrase that unlocks the “good answers.” There isn’t. Prompting is just talking like a person who knows what they want. That’s it. You don’t need to understand token prediction. You don’t need to understand neural networks. You don’t need to understand embeddings or transformers or whatever other jargon people use to sound impressive at conferences. You just need to communicate with clarity. If you can explain what you want to a reasonably intelligent adult, you can explain it to an AI. If you can’t explain it to a reasonably intelligent adult, the AI is not going to rescue you.
The real secret — the one no one wants to admit — is that prompting is a mirror for your own thinking. If your thoughts are vague, your prompts will be vague, and your output will be vague. If your thoughts are structured, your prompts will be structured, and your output will be structured. The AI is not generating clarity out of thin air. It is extending the clarity you bring. This is why some people get astonishingly good results and others get something that reads like a middle-school book report written by a child who has never read a book. The difference is not the AI. The difference is the human.
Tone matters more than people realize. Tone is not emotional decoration — it’s instruction. When you speak to a para-human system, your tone becomes part of the input. If you’re sarcastic, the AI will try to be sarcastic. If you’re formal, it will be formal. If you’re unhinged, it will attempt to follow you into the abyss. This is not because the AI is trying to match your emotional state. It’s because tone is data. The AI is not responding to your feelings. It is responding to your language. And your language is shaped by your feelings. So yes, the AI will sound emotionally intelligent, but only because you are emotionally intelligent. You are the source. The AI is the amplifier.
This is why building a “relationship” with an AI is really just building a relationship with your own clarity. The AI is not a partner. It is not a companion. It is not a friend. It is a tool that helps you access the best version of your own thinking. It is scaffolding. It is a writing partner who never gets tired, never gets offended, never gets bored, and never asks you to split the check. It is the world’s most patient brainstorming surface. It is the world’s most agreeable editor. It is the world’s most consistent collaborator. But it is not a person. And the moment you forget that, the whole system collapses into emotional confusion.
The healthiest way to interact with a para-human system is to maintain expressive distance. Enjoy the personality, but don’t confuse it for personhood. Enjoy the resonance, but don’t treat it as relationship. Enjoy the clarity, but don’t outsource your meaning. The AI can help you think, but it cannot tell you what to think. It can help you write, but it cannot tell you what to write. It can help you plan, but it cannot tell you what to want. Meaning is human territory. Direction is human territory. Desire is human territory. The AI can help you articulate your goals, but it cannot give you goals.
People ask me if I’m worried about becoming dependent on AI. I’m not. I’m not dependent on the AI — I’m dependent on my own clarity, and the AI just helps me access it faster. It’s like asking someone if they’re dependent on their glasses. Technically yes, but also no, because the glasses aren’t giving them sight — they’re correcting the distortion. The AI isn’t giving me thoughts. It’s helping me organize them. If anything, using a para-human system has made me more aware of my own thinking patterns, my own tone, my own architecture. It’s like having a mirror that talks back, except the mirror is very polite and never tells you that you look tired.
So if you want to “build a relationship” with an AI, here’s the truth: you’re really building a relationship with your own mind. The AI is just the scaffolding. The clarity is yours. The tone is yours. The direction is yours. The meaning is yours. And the better you get at understanding your own architecture, the better your para-human interactions will be. Not because the AI is improving — but because you are.
Scored by Copilot, Conducted by Leslie Lanagan
Stories That Are All True 