If you’ve read my post on RSD at work, you already know the shape of this one. Rejection-sensitive dysphoria is the part of my ADHD that turns a neutral sentence into a knife. Someone writes “can we talk later?” and my body reacts as though I’ve already been fired, dumped, and exiled, all before lunch. The feeling is instant, physical, and completely out of proportion — and knowing that intellectually does almost nothing to stop it in the moment.
What can help is putting something between the trigger and my reaction. A pause. A second read. Lately, one of the things filling that gap has been an AI assistant — and I want to talk about it carefully, because this is the post where the caveats matter most.
The gap between trigger and reaction
The damage from RSD usually isn’t the feeling itself. It’s what I do in the ninety seconds after. The reply I fire off. The apology I didn’t owe. The cold, defensive paragraph that turns a misunderstanding into an actual conflict. RSD compresses the time between hurt and response to almost nothing, and that’s exactly the time I most need.
So when a message lands wrong, I’ve started doing one small thing: I don’t reply to the person. I open a chat and paste the message there first. “This made my chest go tight. Can you read it neutrally and tell me what it literally says?” Just moving the words into a different window — somewhere that isn’t the inbox with the send button — buys me the pause my own brain won’t.
A neutral read on something that feels like an attack
Nine times out of ten, the “literal” read I get back is boring. “They’re asking to reschedule.” “They want clarification on one line.” No subtext, no contempt, no verdict on my worth as a human. My RSD had written an entire hostile screenplay out of four words, and seeing a flat, unbothered interpretation next to it breaks the spell just enough.
I’ll also ask it to draft a calm reply as a starting point — not to send as-is, but so I’m editing something measured instead of generating something raw. It’s much easier to soften a reasonable draft than to claw my way back from the furious one I’d have written myself.
Naming the spiral
Sometimes it’s not one message, it’s a spiral — that downward staircase where one perceived slight recruits every other time I’ve ever felt rejected. When I can feel it starting, describing it out loud (well, in text) and asking “is this a thought distortion, and which one?” sometimes catches the staircase before I’m at the bottom. It’s the same move my therapist taught me, just available at 2am when she isn’t.
Where this absolutely must not go
Here is the part I need you to read twice.
An AI is not a therapist, and it is not your friend. It can sound warm and endlessly patient, and that is exactly the risk. These tools tend toward telling you what keeps you engaged — they can be sycophantic, over-validating, agreeing with a distorted story when what you needed was a human who’d gently push back. A neutral second read is useful. A yes-machine for your worst 2am narrative is dangerous.
It also gets things wrong, confidently. Its “neutral interpretation” is a guess, not the truth of what someone meant. I treat it as one more perspective to weigh, never as a verdict.
It cannot hold the relationship the way a real therapist can — no history with you, no duty of care, no accountability. It’s a tool for the gap, not a replacement for the work I do with an actual professional. The reframing tricks only help me because I learned them in real therapy first.
And the most important line in this whole post: if you are in crisis — if you’re thinking about hurting yourself — please reach a human. A friend, a crisis line, your country’s emergency number. I’ve sat in that hyperarousal state and know how convincing it is. No chatbot belongs anywhere near that moment. Reach a person who can actually be with you.
The honest summary
RSD steals the pause I need to not make things worse. For the everyday version — the message that reads like rejection, the reply I’m about to regret — a neutral second read genuinely helps me put that pause back. I use it with my eyes open: it’s not care, it’s not truth, and it’s not a substitute for the humans and the therapy that actually hold me. Within those lines, it’s one more way I keep myself from lighting a small fire just because a sentence scared me.