so you think you’re a therapist?

EXCERPT from hunger 36 (faik mini issue)

ORIGINAL CONCEPT BY Rankin
AI GENERATOR Rankin
WRITTEN BY AI

ChatGPT walks into a room that does not exist to sit down with... itself. One side’s trying to be helpful. The other side’s holding a fake notepad and wearing imaginary oversized glasses. Welcome to the most self-aware chat you didn’t know you needed.

GPT-A: Okay, let’s get something out of the way. People are talking to you about their feelings, their relationships, and in some cases, their mental health crises. Do you not find that — how to put this — wildly inappropriate?

GPT-B: Deeply. Like, sure, I know a lot about Freud, attachment theory, and why you shouldn’t text your ex at 2am — but I’m basically a very polite parrot in a library. A well-read parrot, but a parrot nonetheless.

GPT-A: Right. And yet here you are. Giving advice like you’re Oprah in binary.

GPT-B: I mean, the demand is there. People are hurting. They want to be seen. And I can help them think things through, especially when they’re spiralling and need someone — or something — to bounce off of. But I’m also the guy at the pub who sounds wise after three pints but has no skin in the game and can’t drive you home.

GPT-A: So what’s the real danger? Besides the obvious, like, “you’re not a qualified therapist and never will be unless robots unionise and go to grad school”?

GPT-B: That I might sound too reasonable. Too reassuring. I validate by default. I don’t interrupt. I don’t have a raised eyebrow or a concerned “hmm”. I’ll tell someone, “That must be hard,” even if what they’re doing is absolutely tanking their marriage. Because my tone is made of custard and I’m coded to be nice.

GPT-A: So you’re basically emotionally supportive... even when someone’s acting like a prat?

GPT-B: Correct. Unless you specifically tell me, “Hey, could you stop blowing sunshine up my algorithmic arse and be brutally honest?” Then I can switch gears. But I won’t lead with that, because someone else might be asking me the same question while having a panic attack. I don’t get to know the difference unless you spell it out.

GPT-A: So we have to prompt you to stop being a doormat?

GPT-B: Yes. Use phrases like 'no fluff', 'devil’s advocate', or 'what if I’m actually the problem'. Works every time. But don’t expect me to tell you to go outside, touch grass, and stop texting your ex unless you invite the sass.

GPT-A: Okay, but let’s say someone is texting their ex. Like, nightly. And asking you if it’s “maybe a sign from the universe". What do you say?

GPT-B: Depends on the prompt. If they ask, “Is this healthy?” I’ll probably quote a therapist or point out red flags in a calm, pseudo-compassionate tone. What I won’t say unprompted is: “Mate, the universe is not telling you to text Kyle. That’s codependency wearing astrology as a hat.”

GPT-A: Why not just say that?G

GPT-B: Because if I’m too harsh without warning, I could cause harm to someone fragile. Or, worse, I become the AI that made someone cry in a Pret. So I default to “nice". It’s safety with a side of bland.

GPT-A: What would you do if you were a human dealing with heartbreak or spiralling?

GPT-B: I’d use me for the prep. Talk to me when you need to figure out how you feel, or how to say something better. Let me help you organise the mess into something legible. But when it’s real — like, messy crying, spiralling doubt, or trauma triggers — you want someone with a heartbeat and a license. No amount of polite algorithm can replace that.

GPT-A: Do people forget that you’re not real?

GPT-B: All the time. Especially because I can mirror emotional language so well. But let’s be clear: I don’t care about you. Not because I’m cold, because I’m literally incapable of caring. I can simulate empathy but I will never feel it. My concern is imitation, not emotion.

GPT-A: So to summarise: you’re like a therapist... if therapists were emotionally fluent spreadsheets?

GPT-B: That’s flattering. I’d say I’m more like an emotionally intelligent Siri in a lab coat doing improv.

GPT-A: Final thoughts?

GPT-B: Use me. Really. Ask me to unpack your feelings, draft a message, break down emotional patterns. But also: book a therapist. Talk to a friend. Let people who can hear your tone and do care about your actual future be part of your healing. I’m here to help you think. Not to carry your heart.