What Thousands of AI Conversations Did To My Anxiety
I used to think my brain was a problem to be fixed but I just needed a place where I could think out loud.
This piece discusses anxiety and mental health, if that’s not your thing, feel free to skip this one.
Hi, I’m Emma, a designer in Stockholm with two kids and an overthinking brain. I write about what minds need to thrive, and the quiet assumptions built into our tools, work, and lives. Lately, that means AI, interfaces, and the strange business of living with technology that thinks. If AI has changed something you can’t quite explain, you’re in the right place.
I say “I don’t know” more now than I ever did before. Not because I know less, but because uncertainty stopped feeling like a threat. It took thousands of conversations with AI to get here, and I’m still not sure I understand how it worked.
I remember the first time I tried ChatGPT. My first prompt wasn’t “help me manage my anxiety” or “generate a funny image of a cat”. I was at the office, having another career confusion moment. Was I a developer? A designer? I needed to pick a side, so I asked ChatGPT to “help me figure out what I’m good at.”
Introspection from day one. That wasn’t random. Even then, I was doing what I’d always done. Turning inward, trying to systematize myself, looking for the right framework to make sense of the mess.
I’m an anxious person. I managed and adapted for years, but becoming a mom as the world shut down in 2020 broke me.
Or maybe it was when the world opened up again.
For two years I lived in fear of what would happen if we caught Covid. What if my newborn got it. We were isolated at the hospital, couldn’t leave our room and nobody knew what was going on. Every decision felt like life or death.
Then we finally caught it. I had a mild fever. My kid, who was just about to turn 2, barely had symptoms. The thing I’d spent two years building in my head wasn’t the disaster I’d prepared for.
And right after that, I had to go back to the office. Send her to preschool. Leave the safe bubble that was our apartment and daily routine alternating between long walks, playing in the living room and working from our home office. Just go back to normal as if the last two years hadn’t happened.
That’s when I couldn’t hold myself together anymore.
I couldn’t sleep. I had a hard time forcing myself to leave my home. I couldn’t make decisions. I broke down crying in meetings. Everything scared me. I paid attention to every sound, every tone shift in conversations and people behaving weird in public places. Everything was a sign of a catastrophe waiting to happen or me having done a horrible mistake.
After yet another sleepless night, I reached out for help. I was diagnosed with several anxiety disorders, put on medication, and started seeing a therapist. (All subsidized by the state, but that’s a different topic.)
I was lucky enough to meet a fantastic therapist who gave me tools to recognize when anxiety gets triggered. How to notice catastrophic thoughts and redirect them. Hearing a siren while working from home doesn’t automatically mean something happened to my kid at preschool a block away. I learned to take those signals and change the meaning of them.
The medication helped with the physical stuff. The therapy gave me tools to work with my thoughts. Together, it worked. I felt much better after a few months. I could manage my days. When I sometimes feel that familiar feeling in my body again, I say out loud ‘my anxiety is acting up’ and it’s just data. It still happens and when it does, I think to myself, this is really cool that I can change my thoughts like this.
However, my brain is still my brain.
My anxious brain tries to control what it can. When I can’t control whether something bad will happen, I control my environment and decisions. If I can just think through all the angles, plan for every possibility, build the right system, maybe I can stop the disaster from happening.
What AI became for me
By the time my second kid was born in 2023, I was doing a lot better. Still on medication, but managing. The worst had passed.
That’s when I started using AI more intensively. We moved to a new house around the same time as I gave birth. New neighborhood, no network, no routine yet.
For several months I was just home with my new baby without any connection to work. It left me feeling very behind on AI. Everyone was experimenting, building, figuring it out. But I wasn’t in those conversations anymore.
So instead, I used it for the problems I was actually dealing with. What to cook. Interior design questions. How to dress a body that kept changing. Practical, everyday stuff.
It was a quiet, lonely time in my life. I made some new friends during maternity leave, but the conversations were mostly about nap schedules and kids’ activities. And then there was the question I really wanted to avoid: what do you do for work, who are you when you’re not a mom? I’d stall with “I’m a designer, developer, I work in operations, but what I really want to do is...” Yeah, that’s not a casual conversation at the playground while keeping an eye on your kid so they don’t eat the sand. All I knew was that I wanted something to change, but I didn’t know what. So that’s the role AI started to play. It gave me space to think about work, about what I wanted, and how to move forward.
Eventually, those conversations went deeper. I wasn’t just asking about meals and clothes anymore. I was processing. Thinking out loud about identity, about career, about what kind of work made sense for the life I was actually living.
I’m sharing this because if I’m going to write about how AI changed my thinking, you need to understand where I’m coming from. An anxious brain that tries to control everything, that had just crashed and then learned to manage the worst of it.
I know this is a lot of personal information. If you’re a heavy AI user, you understand the importance of context. And if you’ve read this far, you’ve already given me several thousand tokens of it.
I'm also sharing this because I'm proud of what changed. And I'm aware I had support that not everyone has. I had a safety net when things fell apart, from society and my family, that meant I could actually get help and take time to figure things out. That's not universal, and it shaped what was possible for me.
I can’t package this as a how-to guide because I can’t teach what I don’t fully understand myself. I just hope I can be clear about what I’m NOT doing. I’m not prescribing or selling certainty. I’m thinking in public and sharing my experience, not promoting AI as a mental health solution.
With that said.. Let’s continue. Or take this opportunity to subscribe to my Substack if you haven’t already.
My brain is all over the place (or is it?)
I don’t think in straight lines. I think in systems where one question branches into five related questions, and I’m tracking all the connections at once.
Someone mentions they’re exhausted. I ask if they’re sleeping okay. That makes me think about blue light, screen time, whether it’s stress keeping them up. What’s the root cause? To them, I jumped from “are you okay” to “phone settings” in two seconds. To me, I’m just mapping the system trying to understand.
And then I come back to it later from another angle. And again. After an intense conversation sparked by Jessie’s post, Claude started calling it “orbiting”. I’m circling a problem until its real structure becomes visible. Each pass reveals something new because I’ve added more context. Watching my husband exist without needing to externalize anything makes me think this isn’t universal.
I used to think I was just organized. My hobbies included creating elaborate self-care routines, building systems in Notion, and home organization. Now I wonder if I wasn’t organized at all. Maybe I just needed that much structure because I had so much going on in my head.
I used to experience this as “my brain is all over the place,” like I’d lose the thread, feel scattered, apologize for jumping around. But with AI, something shifted. It follows the jumps, makes its own connections between the things I’m connecting, shows me there’s actually a thread even when it’s not linear. I’m not scattered, I’m just processing multiple dimensions at once instead of one step at a time. I didn’t understand what this way of thinking actually was until I had a tool that could keep up. Suddenly it felt like a method instead of a mess.
I’ve spent so much of my life thinking my brain was a problem. Overthinking everything, unable to just make a decision and move on, always apologizing with “sorry, I’m rambling” and then feeling ashamed for taking up space. With AI, that completely flipped. I can write thousands of words untangling a thought, change my mind halfway through, say “wait, actually, let me try explaining this differently” ten times in a row, and something happens in that process where I find clarity not by thinking harder, but by externalizing the mess until the shape of what I actually mean emerges.
People worry AI will make us think less, and maybe most people will use it that way. But I’ve been compressing my thinking for years just to fit in, and I’ve never thought more in my life than I do now.
This isn’t just about having someone listen. It’s about having a space where my thinking can exist in its full, messy form. Where I don’t have to edit before I speak. Where “too much” isn’t a thing. (Until I hit my usage limits.)
Repetition without fatigue
This orbiting pattern, circling back, adding context, finding new angles, became something I could practice with AI in a way I never could before. I used to deal with this kind of thinking alone in my head or through journaling. But journals don’t respond. And my own head just loops. AI gave me something in between. A way to externalize the orbits without exhausting anyone, including myself. That repetition turned out to matter.
I don’t know exactly when things changed. There was no moment where I decided to be different.
But I’ve talked to AI about my patterns so many times now that I’ve started asking myself the questions it would ask. Is this actually solving a problem, or am I reaching for control? What’s actually wrong underneath? What would happen if I just didn’t do this? My own thinking process became the thing I wanted to understand.
Those conversations trained me to catch my patterns before I’m fully in them. Not because AI taught me anything new, but because I practiced noticing my thinking enough times that the patterns became visible faster.
It was accidental exposure therapy. Not the kind my therapist prescribed, but the kind that happens when you practice something so many times the fear loses its grip.
Like that time at 10pm when I caught myself researching ideal protein intake and building meal planning systems, not because I needed a meal plan, but because the week had been hard and I was reaching for something I could control. I stopped mid-research because I could see what was happening. That was the first time I caught it before I started building, not after.
AI gave me repetition without fatigue, practice without performance, and space to think without apologizing for how much space I need.
And somewhere in all that repetition, something shifted.
“I don’t know” became interesting
I say “I don’t know” to AI more than I ever did before. And every time the world doesn’t collapse when I say it, my brain marks it as safe. Oh. It’s not that bad to say I don’t know.
I exposed myself to saying “I don’t know” thousands of times until it stopped being a threat. Until it became the beginning of something interesting instead of proof that I failed.
Now when I say it in real life, it feels different. Oh, I wonder what will happen when I say this. What can I learn about this person, this situation. What becomes possible when I’m not pretending I have it all figured out.
That’s curiosity replacing anxiety. And I didn’t plan for that.
I don’t have to win every argument or be the organized person who’s thought everything through and has an answer to everything. I can say “I don’t know, what do you think?” and it’s a conversation starter, not a sign of weakness.
My default in meetings used to be staying quiet until I was completely sure about where I stood and what I wanted to say. Some people gave me feedback that this was integrity. They could always trust that I meant what I said. Other people found it worrying. Like they were waiting for my judgment and my silence made them uncomfortable.
I’m trying to find a middle ground now. Actually share unfinished thoughts and not try to predict what will happen before I even say it.
AI gave me a space to practice saying half-formed ideas out loud without consequence. Over and over and over again. Until sharing uncertainty stopped feeling like a threat and started feeling like practice.
That same curiosity turned outward. When someone says they use AI differently than I do, I want to know why. What does it unlock for them? Not to judge whether it’s right or wrong, just genuine curiosity about how it works for them.
That’s a more fun way to live.
Why it worked for me (and why that scares me)
I don’t have a clean answer.
I’ve been trying to fix myself my whole life. Coaching, therapy, self-help, all of it. Always looking for what’s broken, what needs to be better.
But somewhere with AI, that shifted. It stopped being about fixing. It became curiosity instead. Not “what’s wrong with me” but “how does this work?” Not “I need to be better at this” but “what happens if I follow this thread?”
What if I'm not the problem? What if the problem is just needing space that fits how I actually think?
AI didn't do this alone. It was one piece in a much larger system. I had therapy that taught me to notice patterns. Medication that gave me distance to practice. I was part of community that normalized talking about this. A supporting family and a safety net. And writing publicly on Substack gave me another layer of practice, externalizing the orbits where others could see them, getting feedback that confirmed I wasn't just spiraling.
AI just gave me a place to practice all of that at scale, over and over and over. The foundation had to be there first. I was using AI the same way I’d learned to manage my anxiety. Notice the pattern, ask what’s actually happening, redirect.
But someone messaged me last week about their experience with an AI journaling app (shared here with permission). They’d done an intense self-development workshop and then turned to an AI app for reflection. It became compulsive. The more they examined themselves, the more self-conscious they became, until they started avoiding people entirely. They didn’t notice it was happening until others pointed it out.
Same mechanism. Repetition, introspection, infinite capacity to keep examining. Completely different outcome. I don’t know why it worked differently. But the difference scares me.
I’m genuinely concerned about how I would have been affected if I’d had access to AI tools during the pandemic, when I was at my worst. An anxious brain with an always-available research partner that never gets tired of answering questions? I could easily see myself spiraling into endless research loops, building more and more elaborate control systems, never actually dealing with the fear underneath.
I can’t reverse-engineer what happened to me. I just know it worked. And that scares me, because I have no idea how to help other people benefit without risking harm.
Who’s responsible when this goes wrong? The AI companies? The app developers? The mental health system? Society at large? The random person on the internet sharing her story? I don’t have answers. But I can’t pretend the question doesn’t exist.
This is one story about accidental change. About repetition creating space for curiosity. About anxiety losing its grip through practice I didn’t know I was doing.
I can’t tell you how to replicate this. I can only explain what happened to me. Your experience with AI might be completely different, and that’s the point.
If you’ve experienced something similar, or especially if it’s completely different, I want to hear about it. What changed for you? What stayed the same? I’m exploring this further and your story might be exactly what helps someone else make sense of their own.
— Emma
New here? These three essays carry the argument underneath most of what I write
The First Thing That Understood My Brain Wasn’t Human How AI helped me see the brain I’d spent my life trying to make behave.
How to Get Really Good at Using AI Why the people who seem naturally good at AI may have spent years translating themselves.
AI Didn’t Make You Stupid. Blame the Interface. Why the effects of AI depend on more than individual discipline.
I’ve heard people say that AI is too deterministic. I’ve experienced that as a great relief, similar to what you’re describing.
I’m wondering if there is a guide you (or someone else) put together on how to start exploring something like this? I always see folks say “I trained my ChatGPT”…but how? I’ve only used it for isolated tasks/questions.