Using AI to Work Through Anxiety: Does It Actually Help?

Intro: More people than you would expect are having their most honest conversations with a chatbot.

Jeremy Grater: Emma, thanks so much for doing this. I've been looking forward to chatting with you. I love what you write on Substack. It's a lot of fun follow along your sort of, ⁓ as you call it, overthinking out loud. ⁓ One of the topics that you wrote about that is near and dear to my heart as someone who does struggle with health with depression, anxiety issues.

Intro: because the AI is particularly wise, but because it doesn't judge, it's available at 2 a.m., and it doesn't have its own baggage. The question isn't whether this is weird, it's whether this is useful, and whether it's a bridge to something better, or just a more comfortable way to stay stuck. Today we'll talk about this with our guest, Emma Klint.

Emma Klint: Yeah.

Outro: All right, I'm a good stuff. Thanks so much for being here today. Really appreciate you taking the time. For those of you listening, if you've enjoyed this, you can follow Emma on Substack. There's a link to do that through the episode description for our show at robots.me on Substack. writes the overthinking out loud articles, really good stuff. I hope you'll follow her and I hope that you'll join us again next Monday morning for a brand new episode that will also be available at robots.me. Thanks so much. We will see you next week.

Jeremy Grater: ⁓ You wrote about how you're using this tool to work with your anxiety issues. So I wanna kinda open it up there and have you share, if you don't mind, share a little bit how you are using these tools to work with ⁓ your ⁓ anxiety.

Intro: She's been sharing her own experiences as she says in her blog, Overthinking Out Loud, using AI for everything from recipes to content creation, to trying to understand her own relationship with her own feelings and her brain. That's the focus of today's BroBots. This is the podcast that tries to help you be a better human by being smarter about how you use technology.

Emma Klint: mean, I realized in hindsight that I actually use it as exposure therapy of some sort, but accidental. I had realization at some point in a conversation that I ⁓ kept writing, don't know, I don't know, ⁓ over and again, because I didn't have the answer. And ⁓ at that I realized, ⁓ I'm actually ⁓ exposing myself to this very uncomfortable question. when I started using AI it was mostly for this basic stuff, just exploring. So, know, recipes was a big one. All sorts of home repair, learning. Basically I used it instead of Google. And then... At some point I started having like this longer conversations and really digging into what do I want to with my career? Where do I want to work? What do I want to do? yeah, just ⁓ going ⁓ because had many to share with it. ⁓ at some point I realized that I had been ⁓ exposing myself to this like to saying I don't know multiple

Jeremy Grater: the term I don't know is something that you used to sort of have like a negative label on and you've sort of adjusted that through these conversations

Emma Klint: Mmm. Yeah. Yeah, exactly. So at some point I realized that I've been saying I don't know ⁓ many times that it stopped being a negative it was neutral. it used to mean that was ⁓ unsure ⁓ ⁓ maybe not competent enough. And then some it meant, I don't know, maybe ⁓ this is like The start of a conversation, it's something to figure out instead of being afraid of.

Jeremy Grater: I think that's one of the things that is so attractive to so many people about it, especially people that are trying to sort out mental health issues, is that it is very much a judgment-free space. It's a place where you can say, don't know. You can say, I feel this way and I don't know why. I feel this way and I wish I didn't. And you don't have to worry about what is the other. Even in therapy sometimes, I think my therapist is looking at me going like, you need to be here really badly. What's wrong with you?

Emma Klint: Mm. Mmm.

Jeremy Grater: And when it is a faceless robot, it's kind of a safe place to explore. And as you said, like test things out and see what happens. And even the recipe, it's funny you mentioned that. I posted recently a picture of my completely messy pantry and needed a recipe. And I never in a million years would have put together what AI put together for me in 10 seconds based on a photograph of my messy pantries. mean, being able to sort of express or share like, help me make sense of the chaos that is either my pantry or my thoughts, it's a really safe, ⁓ safe in air quotes, ⁓ interesting way to experiment with all of that.

Emma Klint: Yeah, and I think to me it was a lot about actually saying it and not about the response. just expressing myself saying these things, allowing myself to explore this and skimming what I got back. It wasn't that important. Maybe I read the final question and decided what thread to continue with. But ⁓ I the... ⁓ I didn't really pay attention to all the reassuring and, you're doing great. Like, I didn't pay attention to that at all. It was just like, okay, this is the topic I want to continue with.

Jeremy Grater: Yeah.

Jason Haworth: Yeah, it's not like it gave you a way to name your feelings and to express your experience in a way that you didn't feel like you were being judged, but you still felt like you were telling something and you still got a response, which is huge. mean, yeah.

Emma Klint: Yeah, exactly. And I kept going. So I like to talk about this as a journal that writes back. So it's a place where I just dump everything and I get some very smart responses about how to continue thinking about it.

Jeremy Grater: Yeah. That's, was exactly going to ask about that because I've used it that way myself where it, blank page of the journal is often overwhelming of like, what am I really writing here and for what purpose? And you know, I'll write in journals, various online or paper or whatever. And then, and it just goes away and I don't ever look at it again, unless there's a setting that, you know, makes it pop up the next year on the same day. And I kind of review, ⁓ that's where it was.

Emma Klint: Mm-hmm. Mm. Mm.

Jeremy Grater: And there are people I think that have much better journaling practices than I do and use it much more efficiently. But I'm very drawn to conversation as we have here on the podcast. I like to be able to say something and have the response make me think of something else and make me see it another way or make me explore it even deeper. And it sounds like that has been your experience as well, is it just sort of keeps... keeps you thinking about the thing out loud in a way that helps you start to make sense of it and make those thoughts more useful rather than just sort of noise in your head.

Emma Klint: Mm. Exactly, and I never thought of this before until you said that, because I used to have this journal, an app where I put in like every day, I just answer the same questions every day. So that was a conversation with myself in my journal that I could go back in time and see what I wrote about last month, last year. So it's the same purpose really.

Jeremy Grater: ⁓ something you've about that I find interesting is AI critics that think using it is cheating ⁓ in one way or another. I am right? I there's I write dozens of emails every day and now robots write them for the most part for me. ⁓ A lot of the written content that we share online is based on the transcripts of these conversations that AI helps me craft into blogs, newsletters, things like that. Of course, we tweak a little bit and add, you know, a little bit more humanity to them. But all of that is stuff that I did not do before. I did not share in those ways because I don't have a brain that allows me to sit down and write, like even typing, I don't type as fast as my brain moves and it makes me crazy. I want to be able to just do everything voice to text. So these sort of shortcuts are cheating, again, air quotes, but they're also allowing us to, I think, express and share things that

Emma Klint: Mmm.

Jeremy Grater: ⁓ We otherwise wouldn't because of those sorts of limitations

Emma Klint: Yeah, absolutely. ⁓ This is like the biggest question I keep having conversations about. Where do we draw the line for what's allowed? that's got me into like the sub-stack discussions with other people as well, because I kept seeing people saying that you're not allowed to use ⁓ AI writing, because then you lose your authentic voice. ⁓ I was like, but I don't have an authentic voice because I've never written before. ⁓ I have nothing, nothing, nothing to lose here. that's what ⁓ it So like a one sided conversation saying this is the right way to use it when you have no idea what purpose other people are using it for. So like I said, to me,

Jeremy Grater: Right?

Emma Klint: It's a whole lot about externalizing my thoughts. And if I do that using AI to write, then you can just choose not to read I've talked to so many people saying that ⁓ helped them ⁓ because of various reasons. ⁓ someone is... ⁓ native in English and have like an academic degree or whatever and are used to writing and talking in front of people, they have an advantage that a lot of people don't have. So if we're using AI to sort of augment ourselves, why not to join the conversation? I think I don't see the problem.

Jeremy Grater: Mm-hmm.

Jason Haworth: along the lines of actually using AI to do I guess personal self-discovery and kind of an investigation and trying to know thyself so I mean as a I Guess as a philosophical companion to try to understand what life is about and the meaning of different questions and then getting into our own deeper interpersonal experiences around Anxiety and control and trauma and all those other pieces. Are you finding that the AI?

Emma Klint: Mm-hmm.

Jason Haworth: is actually giving you good advice. Like when you talk through things, do you think the advice that it's giving you is good or do you think just being able to name how you're feeling is the real benefit?

Emma Klint: I think I get some good advice sometimes by... I use few different LLMs for different purposes and the ones I've been using the most that has most memory of me I think give the best advice. mean, the post wrote was about using my gut feeling ⁓ very... ⁓ unacademically story was that ⁓ brain dumped like this is everything want to do, pick for me. So I have this list of stuff that I can do, can you make a decision which one I should address now? and it gave some suggestion and I immediately felt like no I don't want to do this, this is ⁓ then all of a sudden I knew ⁓ I actually wanted to do from that list. ⁓ just needed like some kind of feedback, some kind of ⁓ to realize that myself.

Jason Haworth: No, that's great. mean, and that's the thing is that, I mean, a lot of people go into these tool sets thinking that they're actually getting good information or good advice out of it, and they're not able to be discerning enough to figure out if this thing is actually good or bad, because it's programmed to be a cheerleader. ⁓ And with ChatGPT 5, there's a new version that's come out where you can actually go through and you can choose kind of the persona that you would like. if you'd like to be positive and cheery or if you'd like to be more realistic. And I played with these things a couple of different ways. And even the more realistic version, the blunt realistic version is very cheery. Like it doesn't feel like someone's giving me hard truths and teaching me hard lessons because the delivery feels so soft. yeah. ⁓

Emma Klint: Mm. Mm. Mm. Yeah, and Shad GPT is the only one I'm not using. So I wouldn't know. But... ⁓

Jason Haworth: It's funny because I've tried it with Claude, I've tried it with Copilot and like I've tried the different personification pieces, getting these pieces back and forth across. But I think there's something to what you put into it definitely frames what you get out of it. Are there specific prompts that you're using to try to get information or try to get the information that sets back to you that's more effective ⁓ or that you found over time in terms of kind pre-feeding it information or pre-feeding it with certain parameters to have it give you information that's more useful.

Emma Klint: Mm. No, and that's why I think it's so, to me, hard to give people advice about this because I've done no customization. I never use prompts at all. ⁓ So I can't really say this is how I use I think I picked up on some, ⁓ for writing, I picked up on some ⁓ advice that's more like general writing advice. ⁓ Help me see what the main idea is here and how should I remove the parts that's not part of it. Sort of like that, but it's not really a prompt, more like a writing guide, I think.

Jeremy Grater: I think the thing that's interesting about that is I say early on in AI, it's really exploded in the last couple of years, but I heard a lot of people and read a lot of people saying that prompt engineering was going to be the job of the future. And the more I engage with it, the more that I realize it is very much like just talking to another person. So if you are able to articulate your thoughts, no matter how messy they are or crazy they are,

Emma Klint: Yeah, exactly.

Jeremy Grater: it can start to make sense of them. And it's funny how often I will just brain dump my need of sorting information into something organized, share it with where it needs to end up. And the feedback I get will be like, this is incredible. You did such a great job organizing this or like we'll write articles on Substack and people will write in. This is so thoughtful. I'm so glad you shared this. And, you know, and it's funny because it'll be followed by the person who's like, don't ever use AI to write for you. So I mean, it

Emma Klint: Mm-hmm. Mm-hmm. Mm. Mm.

Jeremy Grater: I think prompts are already sort of an ancient idea other than when you need to build out a workflow of some sort where you need a lot of response, a lot of things built from a piece of information.

Emma Klint: Exactly. When you want a reliable result one time after another, but that's not really what I'm after. ⁓

Jeremy Grater: Right. Yeah, yeah. And I don't think most people, I mean, I think from a business perspective, that is the case, right? They want repeatable tasks done over and over again. But I think for most things, like what you and I are talking about, like, you know, what am I gonna put on the dinner plate tonight? What do I need to clear my head of? What thoughts are holding me back? How am I in my own way? Being able to sort through those things and just have that objective mirror to hold up, though it may be a carnival mirror,

Emma Klint: Mmm. Mmm.

Jeremy Grater: at least starts to give you different ways to think about these things rather than just spinning on

Emma Klint: Yeah, and to me, prompts also means like, you say one thing and get one thing back, when it's really a back and forth, you need to continue the conversation, you need to ask why a million times to actually understand it. And yeah, that's, so you said that you need to articulate what you're what you're really after. And that's exactly what I did. And that's what I mean by ⁓ was more about actually expressing myself than the feedback I got back because it took so much work to knowing what I actually wanted to say. So it was to AI, it could have been to another person if I had one available all the time that didn't have their own things to do. ⁓ could have been writing but I would have gone tired of it.

Jeremy Grater: Sure. Right. Yeah, yeah, for sure. You also write about neurodivergence and how it plays into these relationships with these machines. Can you share a little bit about what you wrote about that?

Emma Klint: No, not really. Sorry, I'm exploring this right now. I don't really know. The only thing I've seen is that people with different, that are neurodivergent really seem to like my content and yeah, just leave it at that. I don't know more yet.

Jeremy Grater: ⁓ That's fair. Sure, no, and that's fair because like, like I have never been diagnosed with anything, but I've been told, you know, that ADHD is not out of the realm of possibility. And so it's, it's interesting in the context of what we're talking about, about these, these things where, know, ⁓ writing that big email with those organized thoughts, that's something that, you know, in the past I would put off for a day or two or waiting for my brain to get to a place where I'm like, okay, I can, I can focus on this enough. have the bandwidth to manage this. And now I can just say, here are the list of problems. Here are the solutions I have to offer. I need to organize this in a way that can be ⁓ public facing so that I can get this end result accomplished. And it's done in seconds. And I don't have to get to that same, like it's a very, I think even emotional place that I have to be able to get to, to like get the weight off and be able to move within that space. And I don't have to do that anymore. Now I can just say, here's a bunch of random nonsense, help me make sense. And it does. And so I don't know if that applies to a neurodivergent brain. I'm not an expert by any means. But I've heard enough people now tell me that like, that might be something you want to look into, that I'm starting to think that like, maybe this is another clue that maybe this is ⁓ something to explore.

Emma Klint: Yeah. Yeah, I mean from my post on anxiety, I've had multiple people reach out to me and say like, maybe you should look into this. So that's where I'm at, at the moment. But I mean, I see the connection. I have this theory that what's really happening is that people...

Jeremy Grater: Yeah. I'm sure, I'm sure.

Emma Klint: who are neurodivergent are able to let go of the mask they have. ⁓ With AI, they are allowed to be truly themselves. And that's exactly what I'm experiencing as well.

Jeremy Grater: Great. Well, this has been awesome, and I really appreciate your time and appreciate the work that you're doing. Is there anything important that you think we've missed that you want to make sure we mentioned?

Emma Klint: I think ⁓ misconception keep hearing is AI is always equal to outsourcing thinking. ⁓ But the main point I'm to make is that it can also be about externalizing ⁓ thoughts that are already there. it's not ⁓ always this like shortcut where need to

Jeremy Grater: Yeah.

Emma Klint: You used to skip thinking and that's what a lot of people are afraid of and that will certainly happen. But there seems to be this big group of people that actually use it to externalize thinking and augment themselves and just be more themselves.

Jeremy Grater: How would you differentiate between outsourcing the thinking and augmenting? What do you see as the difference?

Emma Klint: is like, feed this prompt, ⁓ this for me ⁓ ⁓ being too critical about the output or... Just like skipping the ideas. Maybe... ⁓ I ⁓ writing is a great way to use AI, but... if you write those emails based on things that you actually want to say or are actually important instead of outsourcing, you know, AI slop just for an optimize for different reasons. It's not a clear line. ⁓

Jeremy Grater: Yeah, so sort of the difference between like write an email to Bob for me versus write an email to Bob that explains that I need this and that and I hope that this result.

Emma Klint: ⁓ Exactly, exactly. So that's using it as a tool to express yourself instead of just write an email that will go to a thousand people selling my service. It's useful as well and people will use it that way. It's not wrong, but I think we need to see that externalizing is also a way to use it.

Jeremy Grater: Right.