Using AI When You're Feeling Anxious

We've all done it. Waking up anxious at 2am and pulling up ChatGPT to ask it our burning questions. Are we ok? What can we do to calm down? How do we keep this from happening again? 
Today I'm talking about my own AI use and what I learned from it. I cover the limitations and risks of using chatbots as therapists, and how we can use it thoughtfully to help us grow. Enjoy! 

Get your FREE Fear to Freedom Guide here! 

 TikTok: @heyamandaklay

Instagram: @heyamandaklay

YouTube: @heyamandaklay

Email: amanda@amandaklay.com 

Transcript

SPEAKER_00 0:22

Welcome back to the 18 Minutes Podcast. I'm Amanda, and this is the show where we take everything that is confusing, isolating, and scary about fear and anxiety, and we turn it into something you can actually work with. Today we're talking about something that I haven't seen a lot of people talking about, and I think it's really important. And that is how to use AI when you're feeling anxious. Because let's be honest, we're already doing it. When anxiety hits you out of nowhere, it's so easy to reach for a chatbot and start asking away all of your questions. And I want to talk honestly today about what is helpful about that, what is not helpful, and how we can use AI thoughtfully and in a way that moves us forward towards our goals. As always, I am not a therapist or a medical professional, and everything I share on this show is based on my own experience with fear and anxiety, my journey in overcoming it, and the research that I do for the show. This is not a replacement for therapy. So please only take what's helpful to you, leave what isn't, and listen to your medical and mental health professionals first. A couple quick things before I dive in. I recently redid my website. So it's a lot nicer to look at now. It's a lot more organized. Everything has a place. So if you haven't checked that out yet, you can feel free to go ahead and poke around. All my resources are in one place. Also, the Reclaim Group cohort is coming to an end today, actually, the first cohort as I'm recording this. And the next one is going to start up in August. So if that's something you're interested in, I highly recommend subscribing to my weekly newsletter. You can do that through the website. That's where I announce new resources, new cohorts starting up, and everyone who is subscribed to the newsletter does get early access to signups for the cohort. So sign up for that if you haven't already at amandacley.com. Okay, let's get into it. I want to start by being transparent about my own history with this. I'm talking about AI use because I have a lot of experience with using AI and I am sharing from that experience, not from a place of judgment. I have used AI at 2 a.m. when I'm feeling heart palpitations and I know that it's anxiety, but I type the question anyway because it makes me feel better. So that's where I'm coming from. And really for a long time, I used AI chatbots, kind of like a therapist. I would ask it um questions about my personal life. I would vent to it, kind of using it as like a journal and a journal that talks back, which is amazing at first. I would type out everything I was feeling, everything I would I was scared of, and I would ask it what it thought, and I would ask it, what should I do about my problems? And in the moment, it did feel really helpful. I did get some temporary relief from that. And I know you do too. That's why we keep using it. But after a while, I realized that it wasn't really working for me. It wasn't moving me forward. I didn't find myself growing and changing toward my goals, toward the life that I was trying to live. Um, and if anything, I think it was keeping me stuck a little bit. And so I want to talk about why that happens because I know I'm not the only one who has fallen into this. And I also want to be clear that I actually love these tools. Um, chatbots like ChatGPT, Claude, Gemini, all of them are incredible. They're great tools. I use AI regularly for all kinds of things. Personally, I'm a Claude Gurley, and I think understanding how to use AI is really important as it becomes ubiquitous really in our society. It's becoming a bigger and bigger part of our lives all the time. So this is not an anti-AI episode. Um, this is just a let's be thoughtful about specifically how we use it episode. And in this case, that is in the context of reaching for AI when you're feeling scared or anxious. So I want to start with what AI is actually missing when you use it in this way, because I think it can be really helpful to understand its limitations before we talk about the risks. And maybe you're not aware of its limitations. So I'm gonna start there. And the first thing that AI is missing that's really crucial, is most all of the context about you. AI bots only know what you type into them. That's it. They don't, they can't uh get to know you anymore beyond exactly what you are telling them is true about you, um, unlike a therapist, for example. And in that being said, we don't even fully know ourselves. In my opinion, I don't think we can ever fully know ourselves. We can't see ourselves from the outside. And so um, if you can't see all of your tendencies, all of your patterns, um, then a chatbot isn't seeing those things either. Therapist or even a good friend who has known you for a decade can say something like, Hey, you tend to always do this thing when you're scared. AI can't do that because it wasn't there. It only knows the story that you're telling it right now in this moment, filtered through how you happen to be feeling right now. And that's a big deal. It doesn't have so much context that even a person you don't know very well could get from sitting in the same room as you and talking to you for a while. On that note, um, another thing it's missing is the ability to relate to you. And this is a big one. AI can pull up information and it can put ideas together really well, but it has never felt fear. It's never had a panic attack in a grocery store or sat in the car for 20 minutes trying to get up the courage to go inside. It can describe those things, but what it can't do is offer you the fundamentally grounding experience of sharing with another person who has been through the same thing as you. And the third thing that's connected to that is AI cannot give you the experience of being truly known by another person and being accepted anyway. In my experience, that part of any recovery journey is crucial. There's a lot of guilt, a lot of shame, um, and that's very heavy to carry when we go through something like debilitating fear and being seen by someone else who has also had life experiences that are different from yours and being accepted by them anyway and given grace and patience is truly a healing experience. And of course, we don't get that when we go to AI for relief instead of another person. I think it's also important to understand how these tools uh tend to behave when you bring them anxious questions, because there are a few things that are really worth understanding in helping you make a decision on whether or not to use this in the context of getting relief from anxiety. First, most of these consumer AI products are designed, at least in part, to keep you engaged. That's just how a lot of technology works, from social media to apps to chatbots. What that means is that these tools are not optimized to tell you, hey, you've asked me the same question five times in the last couple days. Maybe you should just close this out and go take a walk. It's not designed to stop the conversation, it's designed to keep responding to you. Another really important behavior of AI chatbots to understand is their tendency to mirror you. AI models are generally designed to be agreeable and validating. There's actually a known tendency in these systems that researchers often call sycophancy. And basically, what that means is that these models lean towards telling you what they think you want to hear. So if you come into a new chat with an AI bot and you're feeling anxious and you're looking for reassurance, that AI bot is very likely going to give you reassurance, whether or not that's the most helpful thing for you in that moment. If you come into a chat with a theory about what's wrong with you, it's likely going to agree with you and validate that theory. Here's why that is so important to understand if you are trying to overcome fear in your life. When you are stuck in a fear pattern, the thing you often need is to not be agreed with. You should ask someone else, I don't know. No, never, because they're not designed to respond that way. The voice and tone is always one of confidence and assurance. But back to the first point, the chatbot doesn't know your medical or mental health history. It's never examined you. It's it doesn't know what questions are the right questions to ask to get to a point where it could tell you one way or another what you need to do going forward. A mental health professional or a medical professional can do those things. And what comes along with that confidence and that authoritative tone often gives us relief because when we have an anxious mindset and we are uncomfortable with whatever uncertainty we're currently feeling, a tone, even if it's just text coming up on a screen, a tone of confidence and surety and authority is going to give us a sense of relief and a sense of grounding, whether or not it's good information. It is very good at putting together cohesive information that flows nicely and that intuitively sounds correct. That does not mean that it is. It could be way off base. And we, if we're going to use AI in this way, we have to be extra, extra careful about our critical thinking in everything we read. And I want to go a little deeper on something I mentioned a minute ago, and that is reassurance. I've done a whole episode on reassurance seeking. So go listen to that one if you haven't yet. But the short version, um, as it applies to AI, is that using a chat bot every time you feel anxious to ask if you're okay, to ask whether a symptom is dangerous, to get told that everything is fine, you're okay, that is reassurance seeking. It's the same pattern as texting your friend, am I okay for the tenth time today, or Googling your symptoms at two in the morning. And the thing about reassurance is that it works for about 10 minutes. You get the answer, you feel a wave of relief, and then the anxiety comes back, often worse, more intense, and you need another, you need another hit of reassurance. Every time you get it, you're teaching your brain that you needed it, that it kept you safe, and that you could not have been okay had you not gotten the reassurance. So using the chat bot every time you feel anxious is not actually solving your problem. It's usually creating a bigger need for reassurance every time you use it. Practicing, not reaching for AI every time you feel a wave of fear, even when you really, really want to, is what actually moves you toward calm. I know that's not the fun answer, and that this is a challenge, believe me. I know, but every single time you feel the urge to open Chat GPT or Claude and ask it if you're okay, and you don't, you teach your brain that you have the capacity to tolerate uncertainty. Doing that over and over again actually builds your freedom. It builds the life you're trying to live, which is one where you are in control of your decisions and you're not hindered by fear. And I touched on this earlier, but I want to come back to it because I think it's really important in this context, but also a com a conversation that I think um is worth having in every context as AI becomes more and more a part of our lives. When you use a chat bot as a sounding board, as someone eventually, as a journal, you are not just risking getting unhelpful information. You're giving up human connection that would have been incredibly valuable. Sharing these experiences and these questions with real people helps you in ways that AI just can't replicate. It gives you accountability for one thing, and it gives you a feeling that you're not alone, which, by the way, you're not. As you share your experiences and talk through your feelings with other people, you will realize that a lot of people are way more anxious than you would ever guess. And it gives you a shared experience. It gives you the opportunity for you or someone else to say, Oh my gosh, me too. I thought I was the only one. By the way, this has been super helpful, super helpful to me, or I stopped doing this and I've grown a lot in the last six months. Being vulnerable with people is one of the best ways to build real connection. And real connection is healing and helps us grow in all kinds of ways. I also have a concern about this on a bigger scale. If we all start routing our deepest, scariest questions through an AI chatbot instead of each other, in the long run, I think that creates a huge problem for everybody. We'll all be walking around, moving through life, pretending even harder than we already do that we're fine while privately we're confessing all of our vulnerabilities to a robot. Imagine the domino effect that has on a society. And we can start as anxious, fearful, scared people to be vulnerable with other people and let other people in to our lives to build connection and to build a pattern not only in ourselves but in other people to choose vulnerability and connection over the convenience of sharing information like that with a machine. Another major issue is confirmation bias on our end. If a chatbot tells you something that you were already thinking might be true, you are very likely going to believe that even more, even when it's wrong, because it's confirming what you already thought. If you are in the thick of some kind of anxiety disorder or just overwhelming fear in general, um, it's very likely that you feel like you need to rest all the time. It's exhausting to feel scared every single day. I know I've been there. I had those tendencies to believe that I was not capable of pushing through fatigue or discomfort. Even more recently, I would use AI to help me organize my to-do list, and I would put in something like, I didn't get half of the stuff done today because I didn't feel like doing it, and so now I need to reorganize it for tomorrow. And the kinds of responses I would get when I input that kind of thing would be no worries. That sounds like a really good idea. You probably needed the rest. Don't feel bad about that. That was exactly the right move. Really uh confirming what I already thought was true. And sometimes that absolutely is the right message. We need breaks, we need to rest. But my AI chatbot does not know when that's the right message and when I need to be challenged on it. A lot of times, what I need is not to take a day off because I'm just not feeling it. A lot of times what I need is to work my discipline muscle and get stuff done even though I'm not feeling like it. The chat bot is going to give me the comfortable answer every single time, whereas a friend, a therapist, or a coach might give me that push that would actually help me grow and accomplish the things that I want to accomplish. Okay, so I've spent a lot of time on the cautions around AI because I think they're really important and not a lot of people are talking about them. But I promised you that this is not an anti-AI episode, and it's not. So I'm gonna tell you how I think these tools can be helpful and when to use them from a grounded place instead of a panicked one. The difference, by the way, is very similar to the principles in my healthy habits episode. It's about your intentions. Using AI from a calm, curious place to help create the life that you're trying to build is great. Using it from a scared and spiraling place to get reassurance is a trap. So you can use the same tool and get a totally different outcome. So, some genuinely great uses for AI. You can ask it for resources on working through fear, book recommendations, podcast recommendations, credible approaches, um, how to find a therapist, those kinds of things. Asking AI to give you or show you resources is a great way to use that tool. You can ask it what methods for whatever specific goals you have in your journey have credible evidence behind them. And then you can have it do a deep dive research session on that method and how exactly it works and why it might work for your specific situation. And then you can bring that information to someone like a therapist, and you can be a little bit more informed about the kinds of things you might want to try. You could have it build a habit tracker to track uh regulating techniques like how much sleep you're getting, if you're eating well, if you're exercising, meditating, those kinds of things. Or even better, tracking how often you did something scary and instead of running from it and avoiding your feelings, you stayed with it and moved through it. I would have loved to have Claude when I was going through my um disordered anxiety recovery because I was doing habit trackers by hand, and then I was going back and looking over um days and weeks and months what I had done, what I was doing now, what the differences were, and now you could have Claude or ChatGPT or whatever AI program build that for you and then have it analyze the data. That is a great way to use AI for fear and anxiety recovery. The pattern here is that AI is a great tool to help you build and plan and learn while you're feeling regulated. It's just not an acceptable substitute for a therapist or a friend or your own hard work and practice at learning to tolerate discomfort. And that's all for today. I hope this was helpful and honestly, I hope um it maybe made you think about your own habits a little bit. If you've been leaning on a chatbot every time you feel a wave of fear, no shame at all. We've all been there, but maybe something. If you're looking for more on this, remember I have an episode dedicated to the topic of reassurance seeking that I think is really helpful, especially if you have been using AI to get reassurance. All of my resources live at amandaclay.com, which again looks a lot harder than it used to, so go check that out if you're looking for it. If you have a question or comment about today's episode, or you have a topic you'd like me to dive into on the show, please send me a contact form on my website. You can email me at amanda at amandaclay.com or DM on TikTok or Instagram atamandaclay. If today's episode was helpful to you, please follow or subscribe for A Listening. And if you have a second, please review. Those are super helpful and other people. Thank you so much for being here, and see you next time.