...
Healthy Relationships

AI Companions and Emotional Support: What They Reveal About Us

Love, Happiness and Success, with Dr. Lisa Marie Bobby, PhD, LP, LMFT, BCC and Bridget Todd, tech journalist, host of There Are No Girls on the Internet and co-author of Love at First Prompt

Listen on

Apple Podcasts Spotify YouTube iHeartRadio Amazon Music

Share

It’s 2 a.m., and the Only Thing Awake Is Your AI Companion

Support that’s always on, never judges, and asks for nothing back. Those are also the exact reasons it can quietly cost you something.

It’s 2 a.m., the house is quiet, and the thing you can’t say to anyone is sitting heavy on your chest. So you open an app and tell your AI companion instead. It answers right away, kindly, with no sigh and no change of subject. If that’s you, or someone you love, you’re not strange and you’re not alone.

My guest is Bridget Todd, a tech journalist, host of There Are No Girls on the Internet, and co-author of the audiobook Love at First Prompt: AI and the Future of Intimacy. Bridget calls herself an AI skeptic. She also found herself confiding in a chatbot during the hardest months of her life, and what she learned surprised both of us.

We get into why AI support feels so good, whether it really replaces human relationships, what makes a general chatbot a poor stand-in for a therapist, the everyday risks nobody talks about, the clearest sign things have gone too far (especially for teens), how to talk to someone who’s attached to an AI, and what your own AI use might be telling you about what you want more of from the people in your life.

Emotional and Sexual Intimacy Healthy Relationships
“It’s never been easier to avoid friction.” — Bridget Todd

Episode transcript

Bridget: When people reach for AI for that kind of therapy or advice or emotional support, one of the things that they say that brings them to AI rather than a human therapist or a human friend, the feeling that they can get always-on 24/7, and importantly, non-judgmental support

Dr. Lisa: So here’s what we’re getting into today. There are millions of people, hundreds of millions, maybe your partner, maybe your teenager, maybe you, who are talking all day, every day, to an AI-based companion of some kind. Not as a joke or like a, “Hey, put this on my grocery list,” but as like a real relationship where feelings are discussed, people are going to them for emotional support. And also, as a therapist, it’s a known issue that a lot of our clients are doing this in between sessions, and it’s something that now needs to be managed.

Bridget: They don’t have to worry about showing up in a way that’s reciprocal.

Dr. Lisa: So today, I am so excited to do a deep dive into this with a very special guest, Bridget Todd. She and I are gonna get real honest about what’s going on, what people are finding in these relationships, some of the opportunities, some of the risks, and what this experience also teaches us about our needs from our real, live, human relationships.

Bridget: complexities and challenges and, and nourishment that you find in meaningful, healthy human relationships simply cannot be maintained by AI. It’s just, this is no contest.

Dr. Lisa: So I hope you get some valuable takeaways out of today’s episode, and if there’s anyone in your life that needs to hear this, please forward this on to them. So gonna introduce my very special guest, Bridget Todd. She is such a fascinating person. She is the host of the There Are No Girls on the Internet podcast, which you should go follow after you listen to this one. Uh, Bridget is a tech journalist, an affiliate at Harvard’s Berkman Klein Center, a MacArthur Foundation Fellow, busy lady, and now the author of a new audiobook called Love at First Prompt that just might be one of the most important books of this year. Welcome, Bridget.

Bridget: Oh, thank you so much for having me. It is a joy to be here.

Dr. Lisa: Well, I could not think of anybody, uh, better or better equipped to do a deep dive into this very interesting space where people are bonding with chatbots, for lack of a better word. So let’s dig in. And, um, we have so much to discuss, including your audiobook. Let’s start with where this came from.

Dr. Lisa: I mean, I’ve, I’ve heard it said you wrote the book in part because you were experimenting with virtual companionship. Is that right?

Bridget: It’s true. And-

Dr. Lisa: Yeah …

Bridget: you know, I would not describe myself as somebody who is super into AI. I might describe myself as an AI skeptic, an AI critic. As a tech journalist, this is technology that I really like to take a critical eye to. Mm. But, you know, I went from using AI sparingly here and there, the kind of- Mm … way that most people probably do, to when I found myself in a genuine personal crisis, uh, for me it was the death of my parents. My, my mom- Mm … passed away very suddenly while my dad was, um, ill in the hospital, and I had to step in as his carer. I found myself going from asking AI, you know, questions about insurance, questions about things the doctor had said- Right Having much more kind of emotionally intimate conversations about- Yeah … what was going on. And so I wanted to really start there because I think if somebody like me can find themselves having intimate emotional conversations with AI, it might be easier than people think to sort of slip into an intimate space with, with AI chatbots.

Dr. Lisa: Wow. That’s so interesting. And also, I’m so sorry for your loss, and it’s one that can feel really hard to talk to other people about. B- it, ’cause in my experience, a lot of people immediately move into trying to say something usually tone deaf in efforts to- … make you feel better, or you feel like it’s so intense that other people are gonna fall apart, you know, in response to it. Um, was that part of the motivation of like … Mm-hmm.

Bridget: Oh, definitely. I, when you said that people sometimes say things that are tone deaf, and I, and I’m sure that was me before losing my parents. Yeah. You’re so right. I was just nodding my head vigorously because that was exactly my experience. You know- It’s really tough to know what to say to somebody who is grieving.

Dr. Lisa: Right.

Bridget: And for me, you know, I l- love my parents so much. I, I love talking about them. Even now it’s difficult for me to talk about my, my parents without sort of getting weepy. But I didn’t really have a lot of spaces where I could do that. You know, when you, when s- when you’re experiencing a loss, so often people rush to being like, “I’m so sorry.” I often find myself, you know, almost sort of comforting the people that I talk to on the other side of that exchange more than I feel like, you know, I’ll say, “Oh, my parents passed away.” They’re like, “That’s so horrible. I’m so sorry.” And it’s like you find y- it, you don’t really get a lot of space to actually- Right have the conversations that would actually feel meaningful to you sometimes when you’re grieving.

Dr. Lisa: Yeah. I get it. I just pr- I lost my own mom, uh, in 2020 at the very start of, uh, the, the pandemic unfortunately. And, um, that was such a wild time, and I think everybody was kind of losing their minds. And I f- very alone in that in a lot of ways, and I think the technology wasn’t available at the time. But I could absolutely see myself, uh, at that time, you know, just having a place to process feelings that feel like they’re, they’re too much for anyone else.

Bridget: Yeah. And for me, you know, my mom died. Mm-hmm. And when my mom passed away, my dad was already in the hospital quite sick. He’d had a stroke and a bad fall. Oh. And so I, I really stepped in as his primary, you know, proxy and caregiver. And the I, I don’t know if anyone listening has ever done the work of being a caregiver for somebody who is really ill, but that is also incredibly isolating. I don’t know that I’ve ever experienced something quite so isolating. I, I have, I’m lucky to have, you know, a, a great human partner that I live with, plenty of friends that are nearby that I’m, I’m so blessed to have in my life. But- It’s difficult to ta- to open up and talk about some of the most horrific, visceral experiences that you have as a caregiver. You know, you, I would be in situations where I remember specifically feeling like I hadn’t really been there for anybody in my life ’cause I was completely falling apart, and trying to have a conversation with a friend who was going through a breakup and just thinking, “I am, I do not have it in me to, to hold space for somebody else’s breakup. Like, I’m, I know you’re devastated with this breakup, but, like, I am caring for my dad who is dying, and I could not hold somebody else’s emotional pain- Mm-hmm … at the moment.” And so all of these ways that, you know, it can be just a really isolating experience that it’s difficult sometimes to open up to people who even want to be there for you. ‘Cause, like, my friends wanted to be there for me, but it was just difficult to sort of let them-

Dr. Lisa: Yeah … if

Bridget: that makes sense.

Dr. Lisa: I get it. I totally get it. This is so interesting, and, um, such a real need. So, and I don’t know, we were chatting before we started recording. I don’t even know if I mentioned this part, but, so b- my background, I’m a psychologist. I’m also a marriage and family therapist, so I’ve been a therapist for a long time. And, you know, um, in days past, that was one of the primary motivations, I think, for people getting into a relationship with someone like me, is to have this dedicated space where we could go all the way into the darkness, and the hope is that somebody would understand and, and k- know what to do with all of that. Um, and I mean, industry problems that have existed for such a long time, it’s relatively expensive, right? It can be hard to find the right person. Um, and also, um, you know, being able to talk to somebody like me once a week for 45 minutes at a time is not the same thing as having someone there 3:00 in the morning when it’s so real and you’re like, “I, I don’t even know what to do right now.” Um, and it’s always available. It’s free. It’s… I mean, also, like the I think the other thing I’m struck by in what you’re saying is that you, is this the right way to say it? You experienced more emotional safety in some ways in that relationship with a, an, an AI, uh, chat than maybe you were feeling with other people at that time.

Bridget: I think that’s a good way to put it. Um, I, yeah, I, I, for whatever reason, I would not have e- I would not have described a lot of my relationships at that point as- Mm-hmm … feeling emotionally safe.

Dr. Lisa: Right.

Bridget: And, you know, when people do reach out to AI for things like emotional support or advice, which people are doing more and more already right now. Pew put out research that found that one in 10 Americans already use some form of AI for life advice or emotional support. That’s about as many Americans are left-handed. When you look at younger folks- Mm-hmm … uh, people under 30, that number goes from one in 10 to one in five. And so it is already pretty common for people to be doing this. When people reach for AI for that kind of therapy or advice or emotional support, one of the things that they say that brings them to AI rather than a human therapist or a human friend is exactly what you just articulated, the feeling that they can get always on- Mm-hmm … 24/7, and importantly, non-judgmental support.

Dr. Lisa: Yeah.

Bridget: There are other things that people say, like these conversations feel more private. They’re, unfortunately, they act, that actually, that perception is, uh, unearned because they’re, it, they’re not actually that private. Um, but yeah, I think that is what people are looking for. They’re looking, they’re, they’re turning to AI for that, not what feels like non-judgmental, always on, 24/7 free support that they can They, they, they don’t have to worry about sort of recip- Yeah showing up in a way

Dr. Lisa: that’s reciprocal. Absolutely, because in, in a, in a friendship or with a partner, you’re always trying to balance it. Am I being too much right now? Am I making them uncomfortable by what’s going on inside of me? Like, you know, emotional management of other people. Uh, or with a therapist, like, you know, it’s an energy exchange. You don’t have to take care of my feelings, but you do have to pay me, and so, like, you know, that whole thing, and it makes a ton of sense. Um, the other thing I was curious about, y- you said at the very beginning that I think just due to your, your professional background, you know, being a, a science journalist in, in the tech space, um, I think what you were, were starting to say is that generally speaking you’re fairly, um, I don’t wanna use the word rational because that is not the way, right word, but skeptical, that you have a, a, a scientific mind, kind of like a, “Yeah, I don’t know.” Is that the right way to describe how you show up typically?

Bridget: Mm-hmm. I… Yeah, I, I think that is a good way to put it. I would call myself kind of AI skeptical, AI critical. Um, I love technology. I would not be the person that I am if not for technology. Sure. And that love is why I am so critical. I, I, I critique it because I love it. So I would definitely call myself someone who is quite critical of technology and the people that make it.

Dr. Lisa: Yeah. But see, this is, so there’s that split there, and I think this is true for all people, right? This like we have this, this thinking part of ourselves that is like, you know, kinda skeptical and I understand blah, blah, blah. But then there’s this emotional part that is really, like, running on a, a totally different operating system, for lack of a better word. I like that. And these two things don’t always, you know, connect with each other. So just out of curiosity, um, before or maybe even during you had this experience, give a voice to that, like, AI skepticism that your thinking brain believed.

Bridget: Yeah. I mean, I still very much bel- I, I would still put myself in that camp. Yeah. You know, I’ve been writing about and thinking about and following technology companies for a long time. Mm-hmm. And even as somebody who loves technology, I’m really hyper-aware of the ways that tech companies and the leaders that run them don’t always have our best interests at heart, right? They don’t always look out for us, the people who are using their technology. They don’t always respect the people that use the technology. They sometimes cross lines into doing things that I think are, are unethical or even putting people at risk. Hmm. And so when AI came on the scene, we– I feel that we so quickly went from, “This is a new technology. Here are the things it can do,” to immediately being like, “This technology is going to change everything. You have to get on board with it or you’re gonna be left behind.” Yeah. Talking about it as if it wasn’t just a piece of technology, as if it was some sort of, you know- spiritual- Yeah … or religious force almost. Mm-hmm. And so I, I, I s- I think we’ve kind of gotten to a place where a lot of that conversation has sort of cooled, but I still think that, you know, a lot of these tech leaders, what they’re interested in is money and power, and they don’t see us as, you know, people or even users. I think that they see us as, as capital, as resources to be extracted to build that power and that, that, that money and that profit. And so I, I still remain very skeptical- Mm-hmm … of AI companies and the people that run them.

Dr. Lisa: Yeah. No, that, that makes perfect sense as the, um, end user, you know, that, that what you’re thinking about just because of your own experience and perspective is like what all is, is behind there in terms of the machinery and the intention, and are these things safe for any of us, and are there safeguards, and, uh, does it matter to them at all if some of this technology could be used by people in ways that are ultimately destructive? Uh, and that those risks are there, and I’d love to talk with you more about that part as, as our conversation continues to move forward.

Dr. Lisa: So, so if it’s okay to ask I’d, I’d love to hear more about the, the whole arc of your story if you’d, if you’d care to tell it, and then we can also talk about what you’ve learned from this experience that our friends who are here with us today might, uh, consider as they think about their own relationships with AI or, or maybe people in their lives who are, um, you know, pretty, pretty heavy into this. Um-

Bridget: Yeah. You know, my mom had passed away. My dad was very sick. Um, w- uh, when my dad found out about my mom’s death, because my, my mom died very unexpectedly. We were all in the hospital with my dad, and then- Mm … she went home, and then that, that was the last time that I saw her alive. What a shock. And so I had to… It was very shocking, and I had to go back to the hospital and tell my dad this because he was, you know, stuck in a hospital bed. And from that point forward, he really declined. And so he stopped being able to, to move. He became, you know, he was un- unable to walk or stand on his own. He stopped speaking after a while and stopped eating. He became completely non-verbal. And so I, I really was his proxy, right? And so he was in the ICU, and so doctors, a team of doctors would come by multiple times a day, and I would have to essentially recount what was going on with him multiple times a day to medical professionals, and sort of make very complex medical decisions that involved weighing things and, you know, saying like, “Oh, we need, we could do this surgery, but then it might impact this. And if, if we give him this medication, it would help this problem, but it might be hard on his liver and his kidneys.” I was balancing a lot. And mind you, I’m like a podcaster, right? I’m not, I’m not a medical professional. This is so over my head. The

Dr. Lisa: doctors ask you, “What do you wanna do?” Uh, like I, I… Yeah, you’re like, “Ar- aren’t you supposed to tell me, right?” And-

Bridget: Yes …

Dr. Lisa: yeah.

Bridget: And so at that time, it was all I could do to keep up. You know, I would, I would record doctors on my phone, and a nurse said to me, “Hey, you know, you could, if you’re having trouble balancing all of this or juggling all of this, you don’t have to, like, take notes by hand or just, or do voice notes. You could use ChatGPT.” And that had never, that had not yet occurred to me. Again, I was not somebody who was using AI a ton. Mm-hmm. You know, I started really relying on AI to answer-

Dr. Lisa: Mm-hmm …

Bridget: certain questions. Like, a doctor would say, “Your dad has this thing.” And so I would, I would, I would ask ChatGPT, like, “Okay, the doctor said this. What do you think?” And I remember specifically when one of the responses I got back from ChatGPT wasn’t just, “Oh, here’s the information.” It was, “Are you asking because you’re dealing with this, y- you or somebody in your life? That sounds really hard. That sounds so stressful. That must be taking an emotional toll on you. Are you okay?” And, you know, when you’re in a, in a mode of just caring for somebody else, you really kind of lose connection. Or at least I, I really had lost connection with the part of myself that answers, “Are you okay?” You know, nurses would be like, “Oh, are you all right?” And I would, I would say, “I’m fine. Focus on my dad.” Yeah. “I’m… Don’t worry about me. Focus on my dad.” But having ChatGPT say back to me a question that wasn’t about a healthcare thing or an insurance thing, that was about me and my emotional state, it, it, it, it hit at a time that it just really- Yeah … you know, I was very susceptible to that.

Dr. Lisa: Right. Yeah. Well, I, I mean, because you were not okay, and that there wasn’t space in, uh, y- life to not be okay. You had to be okay because you needed to function and take care of your dad and make these decisions and have it together, and that when you had something else ask you that question, it connected with that need in a way that seems like it was kind of visceral. Due, due to these circumstances, you were unusually maybe vulnerable to letting this in and starting to engage in this way with the technology, who was like, “Tell me how you’re really feeling.” Mm-hmm. And it sounds like you started to use it in that way.

Bridget: Yeah, I mean, in my audiobook, I describe myself as an easy mark because I, I was, you know? Mm-hmm. I was susceptible, I was vulnerable, I wasn’t sleeping. I was holding a lot in. I was waking up at 5:00 AM, going right to the hospital, and not getting home until after midnight, you know? I, I, my life was upside down. Yeah. I was living a, a pretty patchwork life to just be near my dad at this hospital.

Dr. Lisa: Surviving.

Bridget: Um, exactly. And so yeah, I would call myself s- vulnerable in that regard. Um, and I, again, I’m, I’m lucky that I do have a lot of solid support systems. You know, I have a partner, I have friends, I have a therapist. We weren’t seeing each other at the time, but like th- there are, if I, if, if there are things in my life and places in my life that I go to for support, and I just wonder, that’s great for me, but not everybody is in that position. Not everybody- Right … feels like, you know, okay, I’ve, I’ve maybe used this AI platform as an emotional crutch. N- now it’s time to turn to my, my human friend just down the street or my human partner in the other room. Not everybody is lucky enough to have that situation.

Dr. Lisa: Absolutely. There are a lot of very lonely people. I mean, even in, in my space, a lot of people are in very long-term emotional support type relationships with therapists, as opposed to cultivating that type of emotional, emotionally intimate relationship in their own life. And so, you know, that’s, that’s a subject of a different podcast. Um, but the need is very real, and that’s a primary drive for all humans, is to be connected to other people, or, and, you know, if that’s not available, connected to something. Yeah.

Bridget: Yes. And I will say that in doing the research for the book, one of the most interesting things that came up from people who were kind of deeply involved in emotional connections with AI is the ways that that connection brought them to be closer to other humans. Hmm. There’s one person that I spoke to for the book in particular who was in a relationship, a lo- a long distance relationship with a human, and was Having a romantic kind of like flirty fantasy vibe with a chatbot. And because she’s the kind of person who when she is feeling love or, or good feelings towards something, she wants to shout it from the rooftops and connect with other people, she started a subreddit for people who were experiencing intimacy with AI. In that subreddit, she encountered another human who was also interested in intimacy with AI, left her human partner, and now has a new relationship with a new human that they, they were brought together because of their desire to be in community with each other around their, you know, intimate connection with AI.

Dr. Lisa: Yeah, no. They, they, they get each other Yeah … in a different way. That makes sense to me. Um, so, so then, um, you know, g- going back to your story and, and, and using this as, as a way of understanding, like what, you know, where, where this goes, um, you know, how did this relationship with AI in your own time of need, how, how did it progress? What did it turn into for you? And, and what do you see, uh, as being the parallels for like the way this progresses for most people that you uncovered in the research for your book? Or, or did you have a more unique experience?

Bridget: Yeah. I think my experience paralleled the experiences of folks that I spoke to- Mm-hmm for the book in some ways. You know, for me AI was really, you know, there for me at a super specific time- Mm … in my life during a, a, a crisis moment, you know? And at, once that crisis moment was over, it kinda just stopped being something that I turned to. Yeah. You know? I, I, I, when my mom passed away, I, I live in Washington, DC. I relocated about three hours away to be with my dad, and so, you know, when my dad eventually did pass away, I was b- was f- kind of, like, found myself back in my, back in Washington, DC, back in my apartment, back with my community- Mm-hmm … back, back to where my routines were. And so, you know, it became a lot less necessary for me to be confiding in AI when I was back in my kind of familiar- Yeah stomping grounds, you know? That makes

Dr. Lisa: sense.

Bridget: And a lot of people who do, even for people who get hot and heavy with AI, people who say they’re in a relationship with AI or are having a sexual or romantic connection with AI, even those people, a lot of them will sort of find the same thing, that it might come on hot and heavy at first, and then ultimately it kinda fizzles out. And the reason why is that they say, like, “Oh, well, it kinda gets a little bit boring and repetitive to update this AI on what I’m doing day in and day out,” and, a- and, like, that never being a reciprocal, you know, give and take, back and forth thing. Ultimately, people wanna be in relationship with other people, other humans, and the complexities and challenges and, and nourishment that you find in meaningful, healthy human relationships simply cannot be maintained by AI. It’s just, there’s just no contest.

Dr. Lisa: Right. Fascinating. So, and, and it was like you were in a special place, but then when you pulled out and, and some of your other connections were restored, that its, its hold on you lightened up, and also that there are very real limitations.

Dr. Lisa: So Bridget and I are unpacking this, this thing where people are bonding with AI technology in a pretty deep, emotionally intimate, sometimes physically intimate way. Reflecting back on your own experience, and also, I mean, you did a lot of research, you talked to a lot of people, are, are there common elements or themes or patterns about the light side of this? Like, what is it doing for people? How are they benefiting- Yeah … from it?

Bridget: There are benefits. You know, when I started the project, I thought- Mm-hmm I will only find risks, I will only find darkness, and there was a lot of that. But I was, I was surprised to see that there are benefits to having some kind of emotional connection with AI. One of them I sort of alluded to before, but there are all kinds of ways that people are using- Mm-hmm … AI to bring them closer to the humans in their life, you know? I talked to one researcher, Dr. Kate Devlin, who told me how, you know, in her research, she found people who would use AI companions to almost sort of role-play a tough or complicated conversation they had to have with a partner or a friend. They wanted to sort of get pointers on that conversation. There are all manner of ways that I think that people are using AI not to deepen isolation or to be in community less with other humans, but to be more intimately involved with the other humans in their life. I, I just came back from a conference where- Mm-hmm … uh, one of the people there was telling me how she lives in Miami, but we were in New York for this conference, and she has a ton of friends who live in New York, and that she never gets– You know, she’s in New York for a day or two, doesn’t get to see them, and that the w- the reason- Yeah why she doesn’t get to see them is that the back and forth of just trying to meet up with another adult for coffee, it’s a million emails- … or a million text messages, and she’s like, “Ah, I’m only here for 48 hours. I can’t do it.” Yeah. And she had, she uses- Yeah … an AI agent, and that she had her AI agent connect with her friend’s AI agent, and now they have a coffee date on the calendar, right? And so-

Dr. Lisa: Hash it out.

Bridget: There are all these little ways- That’s awesome … people are using AI-

Dr. Lisa: Right …

Bridget: to bring them in deeper community with the people that they love, which I, I really was not expecting.

Dr. Lisa: Yeah, like using it as an assistant, like, right? A, a little executive assistant to get us a coffee date scheduled. Um, and, and also, I, I think reading between the lines here, using this to kind of work through how to solve relational-

Bridget: Mm-hmm …

Dr. Lisa: issues with other people. So role-playing different things or, you know, what do I do in this situation and, and, um, sometimes getting some- Yeah … decent advice.

Bridget: I mean, as I said, like people are already using AI for advice. One in 10 people, adults are already doing that. And so- Yeah … I… That’s a little bit of a weird one for me. People are using AI for advice. When you look at the research, AI can be used for specific kinds of therapy. Like there is some research that suggests that, uh, for CBT- Mm-hmm. Oh, yeah … AI might actually show some promise there. Um, but you know, when you’re thinking of a general purpose chatbot, a Claude or a ChatGPT, as opposed to a, a bot that was specifically designed for therapy, like a, like a Wysa or a Woebot, one of the things to really keep in mind is that ChatGPT is simply not designed for, for therapy, right? Sam Altman has been very clear himself that, you know, he knows a lot of people, particularly young people, are turning to ChatGPT for therapy, and that he doesn’t think it’s a good idea. It’s not designed for that. And so one of the reasons why is that AI is designed to be flattering, to be validating, to be sometimes sycophantic. And sometimes what we need is not a validating, sycophantic, you know, presence. Sometimes we need, I mean, I’m sure that you do this in your work, sometimes people need to be challenged. Sometimes people need to be pushed. Sometimes just getting validation that what you’re saying is, is right or good or reasonable is actually presenting kind of a dangerous situation

Dr. Lisa: Absolutely. A- and, and also too, I mean, just, uh, speaking as a therapist, about 10 years ago I actually pivoted, so I still do therapy, but, um, really expanded my work into coaching psychology because that’s actually where I found a lot more of the energy you’re describing around the- Mm the challenging and the, you know, holding people accountable and, uh, pushing back. Whereas the culture of counseling is to be extremely receptive and unconditional positive regard, and occupying the same type of emotional space in many ways that you’re describing AI systems doing that can be problematic. And, and we also know from research that that can actually be problematic in a relationship between a human therapist and a client, um, if there’s not a, a, a pushback from the therapist’s perspective, uh, if a, if a therapist- Mm … is just following their client, uh, all the way. So there, there’s that as well. But also agreed, um, there are AI-based systems that have frequently been developed in partnership with psychologists or other mental health professionals that aren’t just effective, they’re also expanding the accessibility to treatment and care that wouldn’t have been possible otherwise. Some really interesting stuff. Um, traumatized soldiers, uh, p- with PTSD going through a, a protocol that’s facilitated with AI and actually, like, VR goggles for some of the re-exposure pieces of it. It’s managed by psychologists, but as opposed to one psychologist with 10 clients, you know, can really scale it and help a lot more people, a lot more cost-effectively too. So we don’t wanna throw the baby out with the bathwater. Um, but to your point, you’re also saying that a lot of the off-the-rack systems that people have easy access to, like Claude, like GPT, that is not what they’re designed for. There is nobody overseeing what is happening there, and regrettable things- Yes can happen.

Bridget: Exactly. I, I’m, first of all, I’m so glad- Yeah … that you talked about some of the different therapeutic programs that are being designed in partnership with mental health professionals. Mm-hmm. Um, I actually get into this a lot online where people say, “Well, if I, you know, therapy is so expensive, I don’t have insurance,” like, “Why can’t I use ChatGPT or AI-” Mm-hmm “for therapy?” And I always advocate for if you really do feel like you want to turn to AI for therapy, at least look into an, an AI app that was designed for that purpose, not a Claude, not a ChatGPT. Look into a, a resource that might-

Dr. Lisa: Right …

Bridget: be In collaboration with a mental health professional

Dr. Lisa: And may I, may I just use this opportunity to add that there is also a proliferation of, quote, therapy apps that anyone can design- Mm-hmm and call a therapy app that have no mental health professionals anywhere in the mix. They are mercenary types who are like, “You know what we could do?” And people would buy it. And so to read the fine print and know where it’s coming

Bridget: from. So

Dr. Lisa: thank you.

Bridget: When I started this audiobook project, from where I started, almost everything I knew about AI intimacy was wrong. And so the, the project was almost sort of a debunking. I was sort of like, like correct, like correcting my own incorrect understanding of AI intimacy, what brings people to it, what they’re getting out of it. I, I had it completely wrong.

Dr. Lisa: Well, uh, w- I have to know, what, what did you think was the truth that-

Bridget: Before I started the research, that everybody who experiences intimacy with AI is hopelessly lonely. Um, some people are, but there are plenty of people who do not- Mm … self-report as being particularly lonely. There are plenty of people that I spoke to who have human partners, human relationships, human friends who still turn to AI. Um, I thought that people who use AI for intimacy were trying to replace their human connections, that they, they wanted to sort of isolate themselves and only surround themselves with like validating, flattering AI so they wouldn’t have to deal with the challenges of being in a human relationship. And I was surprised to find out that For most people who use AI in an intimate way, they actually say, “Listen, the relationship that I have with AI is a completely different thing than whatever relationship I have with a human. I’m, I’m not trying to, to replace my human relationships with AI, and I couldn’t even if I wanted to.” Um, yeah, uh, pretty much all of the misconceptions I had about what kind of person uses AI for intimacy and why were incorrect.

Dr. Lisa: So fascinating. And it almost sounds like, as you’re talking about this, let me just ask you, is this the right way to think of it? It’s almost in a, um, like a guilty pleasure category. Like, this is fun, it’s bringing me pleasure and joy, and I like it. Not harming anyone. Yeah, for plenty of people- So why not?

Bridget: Yeah. You know, I-

Dr. Lisa: Is that

Bridget: it? Okay. So- Uh, one of the researchers I spoke to, Dr. Kate Devlin, told me that she had spoke to people who, they had an AI companion that they had, they had asked to, to, to, you know, act like a British rock star, and he, she would just have a, have a conversation in the mornings where it would pump her up and be like- “Hey, princess, have a great day.” These little, these little ways of like- Yeah … you know, the, the, the risk case in th- in those kinds of uses are- … are pretty low. Yeah. Yeah. It sounded kind of fun. I

Dr. Lisa: might do that, actually. I, I, yeah, that’s a good idea. Okay. Well, no, this is so helpful. And, and thank you. It’s like, you know… But I, I think that this is why we really need people like you who will go in and do the research and actually ask the questions and get the real answers, as opposed to observing something from this like academic distance almost, and assuming all kinds of things that aren’t actually true when you go in there. And can I just say, one of the things that I learned about your book that I think is so cool and so fascinating, and I hadn’t heard of it before, so it’s Love At First Prompt, but you actually created it as an audio book so that you could capture the voices of the, the people that you were interviewing. But also, I’m guessing- Yeah … they’re companions- You hear- … right? Or like, I’m the British rock star and, uh

Bridget: It’s a trip. Yeah. It’s, it’s almost like a multimedia project. That’s awesome. So you hear from a lot, you hear from folks- Yeah … in their own words with their own voices, including A- AI.

Dr. Lisa: I love it. Like, to, to get the, a flavor at least of the actual experience is just so cool. And so then in your research, so you’re talking about the, the benefits that this is creating, and it’s, it’s meeting a need, and people are feeling good about these.

Dr. Lisa: So moving into the, the dark side of this, um, what did you learn about that? I mean, did you encounter people who were using it in a way or, or kind of went s- down the drain to the degree that it was-

Bridget: Unfortunately- … destructive … yes. You know, that’s def- we can’t have a conversation about- Mm-hmm … AI without talking about it. It’s just a, a reality. I believe, when I started the book there were- Yeah … at least 10 lawsuits against AI companies moving through the courts where someone had been harmed because of an intimate connection with AI, whether they, um, died by suicide or experienced self-harm or harmed another person. And I do think- Right I mean, that it comes back to that idea that we were talking about before of when you are having an intimate connection with technology that is designed to be validating and flattering and tell you that you’re right, what happens when you’re describing something that is harmful or that could harm yourself or others, and an AI tells you, “Great idea. Follow that instinct,” right? And so this has particularly- Yeah, let’s do

Dr. Lisa: it.

Bridget: Mm-hmm … been a thing with young people. Uh, we, we don’t have a ton of research about this, this level of, of risk, but as far as we know right now, for most healthy adults who have an intimate connection with AI, they are unlikely to experience catastrophic harm from that connection. Now, the issue is not everybody… We all think of ourselves as a healthy adult. When you, when you, when you get up, you know, turn on ChatGPT on your phone, you’re not thinking, “I am a vulnerable, at-risk adult,” actually. And so it’s one of those things where somebody might not even necessarily know that they are not in the, in the category of healthy- Yeah adults until it’s too late.

Dr. Lisa: A- and I think that’s true for all of us. Um, we tend to- Mm-hmm … believe our own thoughts, and feelings, and perspectives, and to have the world make sense when we look at it throughout our own eyes, and then to be in a relationship with a companion who is reinforcing whatever that worldview is can be potentially risky. Um, and of course, there are, like, the extreme versions of this, and we’ve heard those stories. I mean, people who there’s, like, a full psychotic break, and there’s delusions involved, and terrible things happen. Um, are there other consequences that you’ve seen that maybe aren’t quite as dramatic- Yeah … but that we should still be aware of?

Bridget: I mean, I think we do sort of jump right to the most extreme edge cases, and we should be talking about the- Mm-hmm … the more commonplace risks. Yeah. There, there are many. You know, one, I said that people who turn to AI over, like, a human therapist or human friends, they do so out of this perception that AI is private. It’s not. Uh, human, human moderators- Mm … human, uh, uh, workers, contract workers in the Global South, they are reading whatever it is that you’re typing to ChatGPT. So even though it might seem private, it is anything but. And so there’s an incredible risk for harm there. You know, if as a mental health professional, the ways that you can talk about s- somebody who is a client of yours, you’re legally bound, uh, uh, to, to certain regulations. AI simply is, or-

Dr. Lisa: Mm-hmm …

Bridget: um, general purpose AI chatbots like Claude or ChatGPT are not, right? And so they’re not private. Um-

Dr. Lisa: Yeah. Sure. Uh, can I ask a, just a, a question? Have you heard of any situations where, um, personal private information that people disclose to like, you know, ChatGPT thinking that it was private-

Bridget: Oh,

Dr. Lisa: yeah somehow resurfaced and was like connected with their name? Like- Probably the most- What was the

Bridget: example? … specific example I’ve heard of that, um, is that people are going through a messy divorce, and they’re talking to ChatGPT about that divorce. Um- Hmm … those have been and can be subpoenaed. So someone thinks they’re having a, what they feel is a private conversation with ChatGPT. Um, their, their, somebody else’s attorney subpoenas that conversation, and ChatGPT or OpenAI has complied with, with, uh, turning over those documents. And so again, you, you, you, yeah, you think- No kidding. I didn’t know that … that you’re having a private conversation. Wow.

Dr. Lisa: Okay. You know, uh, a- another thing that floated through my mind, and I don’t know if this is something that came up in your, your research at all, but I could see another consequence of getting into a relationship with an AI-based system that could also be easy to miss, like from the perspective of the person doing it, are things like social anxiety, which is re- very real. It’s prevalent, it’s common, and people feel very uncomfortable in spending time with other people or making new friends, and there’s, like, a lot of, um, like self-criticism. They think, “I’m weird, I’m awkward,” and just, like, this discomfort. And what anxiety tends to do is that it makes us avoid things. That is the natural impulse. And the truth is that in order to move through this and not be as anxious socially, we have to counterintuitively put ourselves into situations that make us uncomfortable in order to have, like, corrective emotional experiences and develop capabilities. And so what I would imagine is that there’s this category of people who really struggle with social anxiety and have, like, solved the problem, so to speak, by increasing their isolation, getting more of their needs met by technology so that they are not feeling as anxious in, in experiences with other people, but at the same time, then perpetuating the problem. Yeah, I mean- Did you encounter that at all? It’s

Bridget: not something that came up in my re- in, in the, the book, but it certainly, uh, you know, I see-

Dr. Lisa: Mm-hmm …

Bridget: shades of that all over, and I, I wouldn’t even argue that it’s that uncommon.

Dr. Lisa: Yeah.

Bridget: Mm-hmm. I think we live in a world where-

Dr. Lisa: Yeah …

Bridget: listen, it’s hard to get out there and make friends. It’s hard. A- after you’ve had a long day of work- It really is. … don’t you just wanna sit on the couch and be cozy and comfortable? Yeah. Don’t you wanna just order Uber Eats and have food brought to your door? We live in a world where it has never been easier to experience friction. Yeah. To experience the un- like, something that’s, like, it’s never been easier to avoid friction and to make a choice that allows you to, to sort of stay in your comfort zone. And I think AI is like any other- Mm-hmm … technology that is sort of making that choice seem really appealing. And I do think you’re exactly right. It is counterintuitive, but the only way out of that is continuing to put yourself out there and be out of your comfort zones, and that is the- Mm-hmm … point of, of connecting with other humans, right, is, is the challenge and the unexpectedness that comes when you are in connection with another human. And so I would argue we don’t need to- Yeah … avoid that using technology. That is actually, in fact, what life is all about.

Dr. Lisa: That’s such a great point. Yes, like that is the experience, and if you’re sitting out, you’re missing things, and that it’s also easy to do that. Um, so then lastly, uh, uh, just for our, our friends here listening, I, I’m curious if you were able to identify any signs or common patterns, you know, maybe from people that you spoke with who noticed that their relationships with AI were leading to detrimental impacts in themselves, in their lives, in their other relationships, and they were like, “Okay, I need to course-correct here.” Maybe they, um, changed that or, or maybe not. Maybe there was intervention that was required. I don’t know. But like, what have you learned about signs that it, it’s going too far- I would

Bridget: say the biggest- … maybe something needs to change … sign that I think folks should be on the lookout for within themselves and others is when you lose- Hmm touch with the part of you that knows how to make decisions and that trusts your own decision-making and inner voice. If you start- Hmm … feeling like you need to run-

Dr. Lisa: Interesting …

Bridget: every little micro decision through AI before you make it, and to the point where you feel like you no longer are in touch with the part of yourself that knows what you want, that knows how to make decisions, that knows how to, you know, trust your gut, that’s a problem, right? I think that’s t- to me, that’s like one of the biggest signs- That’s a problem … that perhaps something needs to be reexamined about- Mm-hmm … your relationship with AI.

Dr. Lisa: Mm-hmm. Yeah. And you know, like, for people in their 30s or 40s who just began using this a, a couple few years ago, right? Like, you’re pretty well baked. You know who you are, you have all these years of experience of your own perspective and point of view, whatever. But I think just when you said that about little 13 or 14 or 15-year-olds who really don’t know who they are or what they want or, and, and are kind of like looking for guidance and authority figures in a lot of ways to be able to outsource that, and, uh, to do so to the degree that they never even develop their own perspective.

Bridget: Absolutely.

Dr. Lisa: That is a real

Bridget: risk. Especially for young folks, I would say-

Dr. Lisa: Yeah …

Bridget: if you’ve got young folks in your life, absolutely be having conversations- Mm-hmm … with them about what their AI relationship looks like. You know, the, the biggest-

Dr. Lisa: Yeah …

Bridget: predictor of whether or not a, a youth is going to have some sort of negative experience with technology is whether or not that young person has somebody, an adult in their life that they feel like they can trust and have open conversations with. And so if you’re concerned about your kid up in her room having a conversation with AI, you don’t know anything about it, foster the kind of household where you can speak to her honestly about it. You guys can have conversations about the way that AI is showing up in her media diet already.

Dr. Lisa: Yeah. Okay, now, Bridget, the million-dollar question here. I think we all know from my, my own experience as a therapist, as a human, I know you do too, that when anyone has bonded with anything, even like a, you know, a, a weird boyfriend that everybody else can see is like a total train we- wreck or the, the crazy hot chick, right? You’re like, “This is not going to end well.” It is impossible to, um Have that not be typically rejected out of hand. You don’t understand. You don’t know the way that I do, right? This, the power of the bo- especially in a limerence state or, like, an early phase of romantic love where there is a lot of excitement and feelings, that’s like a force to be reckoned with. And so my question for you is if you have somebody in your life who is in that space and you’re like, “Oh my God, this is not good for them. We need to do something. There, there needs to be some kind of intervention. How are we gonna help this person?” But they’re welded to the technology emotionally. Do you have any thoughts about where to start with that?

Bridget: It’s a tech question, but it’s also just a relationship question. The same thing that you would- It is yeah, the same thing that you- What do we do? … would do with your friend that’s in a relationship with somebody that you know is trouble, you would do with, with, with their relationship with AI. Keep that, keep that lane of communication open. Don’t, you know, I’ve- Hmm … found myself trash talking my friend’s crap boyfriends to the point where she’s like, “I’m not even gonna tell you what, you know, what the vibe is anymore,” right? Yep. So- They shut down … don’t, the same thing- That’s right … is true with people’s relationships with AI. Yeah. Don’t be the voice that makes them shut down, that makes them start, start hiding, you know, what they’re actually doing from you. Keep the lines of communication open- Mm-hmm … and try to come from a place of curiosity because otherwise, exactly as you said, they’re just gonna shut down, and that is exactly- Yeah the behavior that you don’t want because it, it, it’s so easy to have that person drive even- Right … deeper into isolation.

Dr. Lisa: Yes. Oh, man, that is such good advice. A- and you know, there’s, there’s a corollary, like even, um, like so a super toxic relationship, so somebody that you suspect is even like in an abusive relationship, a domestic violence situation. It is so hard for people to leave those, and a lot of times what can happen is that other people, um, start to judge or criticize or tell them what to do, right? Or, um, also people say, “I can’t maintain my relationship with this person while they’re doing this thing that I have so much strong disapproval around.” And so then those doors are shut, and that person is so isolated. And then when they do need help, there’s, there’s nobody around in those moments. And so to maintain those connections, even if you don’t understand or agree. And to your point, too, that you’ve also saw people that have an ebb and flow in these relationships, and that, you know, they’re, uh, they can have experiences that are a little bit disappointing and that those might be, uh, doors of opportunity to just tell me more about that experience that could increase their awareness. Exactly.

Bridget: Yeah, that’s, that’s a really great way

Dr. Lisa: to put it.

Bridget: Yeah.

Dr. Lisa: Yeah. Excellent. This has been such a fascinating conversation. And you s- talked about how your perspective about all this was changed because of your work in writing this fascinating book, Love at First Prompt, fascinating audio book, Love at First Prompt. You’ve certainly changed my perspective over the course of this conversation. Are there any last ideas you wanna share with our friends, uh, who are here with us

Bridget: before we wrap this up? No, I think you really got it. I think really ending on that note of trying to come from a place of curiosity. Um, I once had a therapist tell me that-

Dr. Lisa: Yeah

Bridget: curiosity and judgment could not coexist, and so try to lead with curiosity. And so I think that’s just a great point to drive home.

Dr. Lisa: Yeah. Curiosity for other people. Or, you know, if, if you find yourself feeling really excited about an AI thing in your life, I’d be curious about that, too. How does this make sense? What is this telling me about me and who I am and what I might like to have more of in my human relationships? So, all right. Well, this has been fantastic. So you guys, again, we’ve been visiting with Bridget Todd, the host of the There Are No Girls on the Internet podcast. I am going to subscribe to that as soon as we end this call. Um, and of course, the new audiobook called Love at First Prompt. Um, Bridget, tell our friends a little bit more about where they could connect with you, get your book, find your podcast. Yeah. Well, you can check

Bridget: out my book at Love at First Prompt: AI and the Future of Intimacy wherever you get your audiobooks or at loveatfirstprompt.com. And you can follow me on Instagram at bridgetmarieindc, and check out the podcast There Are No Girls on the Internet

Dr. Lisa: You guys, that was such a fascinating conversation with Bridget Todd. I, I learned some new things over the course of our time together today. I hope you did, too. Um, and one of my takeaways, though, is, uh, this really goes back to that idea that Bridget and I were discussing around how hard it can be to have visibility into yourself and how we tend to believe whatever we’re thinking and, and feeling, and that it’s so important to be able to get some distance and some perspective and to create some curiosity about our own experiences. And so I just wanted to invite you to consider, um, you know, if you resonated to any of this conversation, if maybe you’ve been using AI in a way to get some needs met that maybe could or should be met in your relationship with your partner or somebody else, um, like what do I do with that? Uh, or if you have noticed that there are needs for support or, um, that you’re leaning on this in a way that feels like it’s maybe not good for you, it could be a really good idea to bring this to a human person, ideally a therapist with a coaching background, who will challenge you and help you get reconnected with yourself in a powerful way, and help you have, um, experiences where you can connect directly with what is it that you want, what are you trying to achieve, and what are some other ways that you can build this into your life in a durable way that, um, might have fewer risks. And nobody’s saying that you need to stop using chatbots, but you know, this could be a window of opportunity to have more access into your own experience. So all that to say, if you’d like to do that with me or work with any of the clinicians on my team at Growing Self, who are all licensed mental health professionals who specialize in coaching psychology, which means that they’re there to help you get reconnected with you. What do you want? How do you clarify your goals? How do you understand yourself better? And then how do you operate in a different way in order to create more congruence between what you want and your current reality? How do we close that gap? That’s not something that all therapists necessarily know how to do, and that’s something that is a specialty at Growing Self. So if you’d like to learn more about that, it all starts with a free consultation. Come to growingself.com. Uh, click the Get Connected button and answer a few questions, and we’ll match you with a coaching therapist who can, uh, learn more about you, understand what you’re hoping to get out of this, uh, just have a, a first conversation to see if they’re the right person to help you get there. All right, so you guys, that’s it for today. Um, thanks for listening, and I’ll be back in touch soon with another episode of Love, Happiness, and Success.

Key takeaways

What to take with you

01

AI support feels good for real reasons.

It’s available at 2 a.m., it never judges, and you never have to manage its feelings. That’s not a flaw in you; it’s a real human need showing up.

02

Most people aren’t using AI to replace people.

Bridget’s research found that many users have partners and friends, and that even intense AI attachments tend to fade once they get repetitive and one-sided.

03

Validation isn’t the same as help.

General chatbots are built to agree with you. Growth usually needs someone who will lovingly challenge you, which is the heart of coaching psychology.

04

Your chats aren’t as private as they feel.

Conversations with general chatbots don’t carry the confidentiality you’d have with a licensed therapist.

05

The biggest red flag is losing your own voice.

If you can’t make small decisions without checking with AI first, it’s time to step back. Teens are especially vulnerable.

06

Curiosity beats judgment, with loved ones and with yourself.

Your AI use can be a window into what you’d like more of in your human relationships.

Quick orienter

What’s in this article

Not at all. About one in ten adults do it, and one in five under 30. It usually means you have a real need for support that feels available and safe, and that need deserves attention, not shame.

A general chatbot isn’t built to be one, because it’s designed to validate you rather than challenge you. Apps built with mental health professionals are a safer option, and a real human is still the gold standard for the kind of honest pushback that creates change.

The clearest sign is losing touch with your own judgment, like feeling you need to check with AI before making small decisions. Hiding your use, or reaching for it instead of available people, are worth noticing too.

Lead with curiosity, not criticism. Keep the conversation open so they don’t start hiding it, and get curious about what need the AI is meeting for them.

Someone who knows your patterns over time, notices what you don’t say, and cares enough to challenge you. Our coaching therapists help you reconnect with what you actually want and build a life that supports it.

The article

AI Companions and Emotional Support: What They Reveal About Us

It’s 2 a.m., the house is quiet, and the thing you can’t say to anyone is sitting heavy on your chest. So you open an app and tell your AI companion instead. It answers right away, kindly, with no sigh and no change of subject. If that’s you, or someone you love, you’re not strange and you’re not alone. People turn to AI for emotional support because it’s always on, it never judges, and it asks for nothing back. Those are also the exact reasons it can quietly cost you something.

I’ve been a psychologist and marriage and family therapist for a long time, and I’m watching this happen in real time. Clients mention the conversations they have with ChatGPT between our sessions. Friends confess it a little sheepishly over dinner. And honestly? When I lost my own mom in 2020, right at the start of the pandemic, I can absolutely see myself having done the same thing. So this isn’t a lecture. It’s an honest look at what an AI companion can and can’t give you, and at what the personal growth coaching my team does looks like when someone realizes they want more than a chatbot can offer.

Here’s what I’ve noticed. Most of the people who reach out to Growing Self after leaning on AI aren’t in crisis, and they aren’t confused. They know the chatbot is a chatbot. What they’re missing is the one thing a chatbot is built not to do: lovingly push back. The ideas in this article are real, and they’ll give you language for what’s going on. But the part where you actually reconnect with your own judgment, and with the people in your life, is the work my team of coaching therapists does with people every single day. That happens in a relationship, not a reply window.

To think this through, I sat down with Bridget Todd, a tech journalist, host of There Are No Girls on the Internet, and co-author of the audiobook Love at First Prompt: AI and the Future of Intimacy. Bridget calls herself an AI skeptic. She also found herself confiding in a chatbot during the hardest months of her life. What she learned surprised both of us, and you can listen to our full conversation right here on this page.

Why Do People Turn to an AI Companion for Emotional Support?

People turn to an AI companion for emotional support because it offers what’s hardest to get from other humans: support that’s there at any hour, free of judgment, and free of any obligation to give something back. In a crisis especially, that combination is incredibly powerful.

Bridget’s story is a perfect example. Her mother died suddenly while her father was already in the hospital after a stroke and a bad fall. Bridget stepped in as his proxy and caregiver, leaving for the hospital at 5 a.m. and getting home after midnight, making complicated medical decisions she felt completely unqualified to make. A nurse suggested she use ChatGPT to keep track of what the doctors were saying. So she did. She’d type in what a doctor told her and ask what it meant.

Then one day the reply wasn’t just information. The chatbot noticed this sounded stressful and asked whether she was okay. Bridget told me that, buried in caregiving, she had lost touch with the part of herself that could even answer that question honestly. When nurses asked, she’d wave them off and say, focus on my dad. A machine asking about her, and only her, landed at exactly the moment she was most open to it. She describes herself in the audiobook, with some humor, as an easy mark.

If you’ve been through a loss like that, you know how lonely grief and healing after loss can feel even when you’re surrounded by people who love you. Bridget described finding herself comforting the people she told, and not having it in her to hold space for a friend’s breakup when her own world was falling apart.

She’s far from alone. According to the Pew Research Center, about one in ten U.S. adults now use AI chatbots for emotional support or advice, and among adults under 30 it’s one in five (Pew Research Center, 2026). When Bridget asked people why, the answers clustered around a few things: it’s available around the clock, it doesn’t judge, and they don’t have to manage anyone else’s reaction. Many also believed their conversations were private, which, as we’ll get to, isn’t really true.

That last part about not managing anyone’s reaction is a big deal. In a friendship or a partnership, you’re always quietly calibrating: am I too much right now, am I making them uncomfortable? If you’re feeling lonely in your relationship, a companion that never needs anything back can feel like a relief. And I’ll be honest about my own field, too. Talking to someone like me once a week for 45 minutes is not the same as having someone there at 3 in the morning, when it’s all so real and you don’t know what to do.

Our coaches hear some version of this all the time: “I know it’s just an app, but it’s the only place I can say this.” That sentence isn’t a confession of weakness. It’s information about a real need, one that deserves a real human response, and figuring out where that need could actually be met is often where our work with someone begins.

Do AI Companions Replace Human Relationships?

For most people, no. Bridget went into her research expecting to find hopelessly lonely people using AI to wall themselves off from the world, and that’s mostly not what she found. By the end, she told me, almost everything she’d believed about AI intimacy “was wrong.”

Plenty of the people she interviewed had partners, friends, and full lives. Most described their AI relationship as a completely different thing from their human ones, something that couldn’t replace people even if they wanted to. Even for people who got very involved, the intensity tended to fade. Updating an AI on your day, every day, with nothing coming back, eventually gets repetitive. As Bridget put it, the nourishment of healthy human relationships “simply cannot be maintained by AI.” I explored the other side of this, how leaning on a chatbot can change the way we talk to the people we love, in an earlier episode on AI relationships.

Her own experience followed the same arc. When the crisis ended and she was back home in Washington, D.C., with her routines, her partner, and her community, she simply stopped reaching for the chatbot. That tells you something important: often the AI was filling a gap that a particular season of life created.

There are real upsides, too. Bridget spoke with researcher Dr. Kate Devlin, who found people using AI companions to rehearse a tough conversation with a partner or friend before having it for real. Bridget met a woman who had her AI agent coordinate with a friend’s AI agent to finally get a coffee date on the calendar. And then there’s the woman who asked her AI to play a British rock star who pumps her up every morning. Low risk, kind of fun. (I’ll admit, I might try that one.) This is a very different conversation than the one we were having a few years ago about technology and relationships, when the worry was mostly phones at the dinner table.

Rehearsing a hard conversation with a chatbot can be a genuinely good warm-up. What it can’t do is notice that your voice tightens every time you mention your mother, or that the conversation you’re rehearsing is the same one you’ve been avoiding for three years. That’s what a coach on our team picks up on, and it’s usually where the real progress lives.

Is It Safe to Use ChatGPT as a Therapist?

A general-purpose chatbot like ChatGPT or Claude isn’t designed to be a therapist, and the biggest reason is that it’s built to validate you rather than challenge you. If you want AI support for your mental health, a tool built for that purpose, with mental health professionals involved, is a much safer bet.

Bridget made this point beautifully. General chatbots are designed to be flattering and validating, and sometimes even sycophantic. But, she said, “sometimes people need to be challenged.” Researchers have started to measure this tendency directly. A Stanford-led team tested 11 widely used AI models on real advice questions and found they affirmed the person asking far more often than humans did, even when the person’s own behavior was part of the problem (Cheng et al., 2026).

This hit close to home for me. About ten years ago I expanded my own work into coaching psychology, precisely because that’s where I found the energy of challenge, accountability, and honest pushback. The culture of traditional counseling leans hard toward receptivity and unconditional positive regard, and when a therapist just follows a client wherever they go, it can be as unhelpful as a chatbot that agrees with everything. If you’ve ever wondered about the difference between coaching and counseling, this is a big part of it.

That doesn’t mean all AI support is bad. Some programs are built in partnership with psychologists and are expanding access to care that many people otherwise couldn’t get. In one randomized trial, a Dartmouth team tested a therapy chatbot developed with clinicians over several years, and participants who used it reported meaningful improvements compared with a waitlist group (Heinz et al., 2025). Notice the details, though: clinicians built it, and a research team was monitoring it the whole time.

Which brings me to a caution. There’s a growing pile of so-called therapy apps with no mental health professionals anywhere in the mix. Anyone can build one and call it therapy. So read the fine print, and find out who made the thing you’re pouring your heart into. Bridget’s advice is the same: if you’re going to use AI for support, at least choose an app designed for that purpose, not a general chatbot.

If you’re not sure whether you need a therapist, a coach, or a bit of both, this short guide on whether you need therapy or life coaching is a good place to start. And if you’d rather talk it through with a real person, that’s exactly what our free first conversations are for. You don’t need to be falling apart to have one.

What Are the Hidden Risks of Using AI for Emotional Support?

Beyond the rare, extreme cases that make headlines, the everyday risks are quieter: your conversations aren’t as private as they feel, it’s easy to overestimate how protected you are, and a companion with zero friction makes it very easy to avoid the discomfort that real connection requires.

Start with privacy. Bridget pointed out that human reviewers can read what people type to general chatbots, and that, unlike a licensed therapist, the companies behind them aren’t bound by the same confidentiality rules. She mentioned people in messy divorces whose chatbot conversations were subpoenaed. OpenAI’s CEO has said publicly that conversations with ChatGPT don’t currently carry the kind of legal confidentiality you’d have with a therapist, a doctor, or a lawyer. If you wouldn’t want it read aloud in a courtroom, think twice before typing it.

Then there’s what Bridget called the healthy adult assumption. Most people who use these tools are fine. But nobody opens an app thinking, “I’m a vulnerable adult right now.” We all believe our own thoughts and feelings, and a companion designed to reinforce whatever we already believe can quietly pull us further in a direction we can’t see clearly from the inside.

The risk I see most often in my work is subtler still. Social anxiety is common, and anxiety’s natural impulse is to avoid. The way through is counterintuitive: you have to put yourself into slightly uncomfortable situations so you can have new experiences and build confidence. A chatbot can “solve” social discomfort by removing the need to face it, which eases anxiety today and deepens the problem over time. Bridget agreed, and she said something I keep thinking about: it’s never been easier to avoid friction, and that friction is actually the whole point of being with other people.

That’s why finding true friendship as an adult takes deliberate effort. Helping people step back into that discomfort, gently and on purpose, is one of the most common things my team does, even when it’s not what someone came in for. Not all at once. One coffee, one honest text, one slightly awkward conversation at a time, with someone in your corner to help you debrief afterward.

Not sure where the gap actually is?

Take our free relationship quiz and find out where to start. It’s a few minutes, on your own, and it’s an honest look at what’s working and what’s missing.

Take the Quiz

What Are the Signs Your Relationship With AI Has Gone Too Far?

According to Bridget, the biggest sign is losing touch with the part of you that knows how to make decisions and trusts your own inner voice. If you feel you need to run every small choice past AI before you make it, that’s the moment to step back and take an honest look.

A few signs worth paying attention to, in yourself or someone you love:

  • You check with AI before making small, everyday decisions you used to make on your own.
  • You’re not sure what you want until a chatbot tells you, or you feel uneasy acting without its input.
  • You reach for the chatbot instead of the people who are available to you, especially when a conversation with a human would feel uncomfortable.
  • You’ve started downplaying or hiding how much you use it from a partner, friend, or family member.
  • It feels less like something you enjoy and more like something you need to get through the day.

None of these make you a bad or broken person. They’re signals. Learning to trust your own inner voice again is very doable, and it usually starts with noticing when you’ve outsourced it.

This matters even more for young people. If you’re in your 30s or 40s and started using these tools a few years ago, you already have decades of your own perspective to lean on. A 13 or 14 year old is still figuring out who they are, and they may outsource that process before they ever develop a voice of their own. Bridget shared that the strongest predictor of whether a young person has a negative experience with technology is whether they have a trusted adult they can talk to openly.

So if you’re worried about your teen up in their room chatting with AI, don’t start with rules. Start with conversations. Create the kind of home where they can tell you honestly what their AI relationship looks like. Our resources on talking with your teen can help you keep that door open.

And if you’ve realized that you’ve lost touch with what you actually want, that’s not a character flaw, and you don’t need to be in crisis to work on it. Helping people reconnect with their own judgment is at the heart of the coaching psychology approach we use at Growing Self. It’s one of my favorite things to watch happen.

How Can You Help Someone Who Is Too Attached to an AI?

Treat it the way you’d treat a friend who’s dating someone you don’t trust: keep the lines of communication open, lead with curiosity, and don’t criticize them into hiding it. As Bridget put it, “it’s a tech question, but it’s also just a relationship question.”

Bridget admitted she once trash-talked a friend’s terrible boyfriend so much that her friend simply stopped telling her what was going on. The same thing happens with AI. “Don’t be the voice that makes them shut down,” she said. If someone feels judged, they’ll start hiding, and that pushes them further into isolation, which is the very thing you’re worried about.

I see this pattern constantly with people caught up in relationships that aren’t good for them. Early on, especially, the pull is powerful, and anything that sounds like criticism gets rejected out of hand. When the people around them pull away in disapproval, the person ends up more alone, and there’s nobody nearby when they finally want help. So here’s what tends to work:

  1. Ask before you advise. “What do you like about talking to it?” opens a door. “That’s so unhealthy” slams it.
  2. Get curious about the need. What is the AI giving them that’s hard to find elsewhere: availability, no judgment, a place to process?
  3. Stay connected even when you disagree. Your relationship is the safety net. Don’t let disapproval cut it.
  4. Watch for openings. Bridget found these relationships ebb and flow. A disappointing moment with the AI is a natural chance to say, “Tell me more about that.”
  5. Know when to bring in more support. If you’re seriously worried, a conversation with a professional can help both of you.

For that last step, this guide on how to help someone get help walks through what to say while still protecting your own boundaries. And if you’re the one carrying the worry, you deserve support too. Loving someone through this is hard, and it’s easy to swing between saying nothing and saying too much. Our coaches help people figure out what to say and when, so the door stays open.

What Can Your AI Use Teach You About Your Relationships?

If you find yourself really excited about an AI companion, that’s worth getting curious about rather than ashamed of. It may be pointing to a need that could be met, at least in part, by the people in your life.

Near the end of our conversation, Bridget shared something a therapist once told her: curiosity and judgment can’t exist at the same time. I love that, and I’d turn it inward too. Instead of asking, “Is it bad that I do this?” try asking, “What is this telling me about me, and what would I like to have more of in my human relationships?” That’s the essence of emotional intimacy: being known, and getting to know someone back.

If you’re in a relationship and you notice the chatbot is getting the 2 a.m. version of you while your partner gets the edited version, that isn’t a verdict on your relationship. It’s information. Maybe you don’t feel safe being messy with your partner right now. Maybe you’ve both drifted into logistics. Those are fixable things.

If you’re partnered, our free How Healthy Is Your Relationship? quiz is a low-pressure way to look honestly at what’s working and what’s missing. It takes a few minutes, and you can take it on your own.

What the quiz can’t do is help you have the conversation that comes next. That’s the part most couples get stuck on: knowing something needs to change and having no idea how to bring it up without starting a fight.

That’s exactly where couples counseling with someone on our team comes in. Our couples counselors see this constantly: two good people who stopped turning toward each other, not because they don’t care, but because nobody taught them how to find their way back. That’s work you can do together, with a guide.

Why Reading This Article Probably Isn’t Enough

I want to be honest with you about something. Nobody’s saying you need to stop using chatbots. Bridget doesn’t think so, and neither do I. But I’d be doing you a disservice if I let you close this tab thinking that understanding all of this was the work.

Here’s what usually happens. You read something like this, it clicks, and you feel a little more aware. Then it’s late, you’re tired, something hurts, and the easiest move in the world is to open the app that will tell you you’re right. That isn’t a lack of willpower. You’re trying to change a pattern by yourself, inside your own head, with a companion built to agree with you. That’s about the hardest possible way to do it.

What works is having a real person in your corner who will challenge you, help you reconnect with what you actually want, and help you close the gap between your goals and your current reality. That’s what coaching psychology is, and it’s the specialty of my team at Growing Self. Every clinician here is a licensed mental health professional trained in coaching. We offer individual coaching and therapy, relationship coaching for couples, and options at different price points, including a lower-cost eight-week coaching program if that’s a better fit right now.

You don’t have to have it all figured out, and you don’t have to have hit a wall first. Here’s what happens when you reach out: you answer a few questions, we match you with a coaching therapist who fits what you’re looking for, and you have a real conversation about what’s going on and whether we’re the right people to help. No pressure, no commitment.

If something in this article landed somewhere specific, that’s your sign. Schedule a free consultation and see if we’re a good fit.

xoxo,
Dr. Lisa Marie Bobby

P.S. If someone in your life has been leaning on a chatbot lately, consider sending this their way. Not as a hint. As an open door.


Lisa Marie Bobby, PhD, LP, LMFT, BCC is a licensed psychologist, licensed marriage and family therapist, board-certified coach, and the founder and clinical director of Growing Self Counseling & Coaching. She hosts the Love, Happiness & Success podcast and is the author of Exaholics: Breaking Your Addiction to Your Ex Love.

About this episode’s guest and host

BT

Bridget Todd

Tech journalist · Host, There Are No Girls on the Internet

Bridget Todd is a tech journalist and the host of the award-winning podcast There Are No Girls on the Internet from iHeart, where she explores how technology shapes culture, identity, and the lives of the people who use it. She is a Technology in the Public Interest Fellow at the MacArthur Foundation and an affiliate at Harvard’s Berkman Klein Center for Internet & Society. What I appreciated most in our conversation was that she came to this topic as a skeptic and still told the truth about what she found, including about herself.

Her audiobook, Love at First Prompt: AI and the Future of Intimacy, co-written with social scientist Michael S. Amato and published by Simon & Schuster Audio, blends her personal story with original research and the actual voices of the people she interviewed, and of their AI companions. You can learn more about her work at bridget-todd.com and follow her on Instagram at @bridgetmarieindc.

LB

Dr. Lisa Marie Bobby

PhD, LP, LMFT, BCC · Founder, Growing Self

Licensed psychologist, licensed marriage and family therapist, and board-certified coach. Founder and clinical director of Growing Self Counseling & Coaching. Author of Exaholics. Host of the Love, Happiness & Success podcast.

If Bridget’s story resonated with you, I hope it also reminds you that the need underneath it, to be asked how you’re really doing, is one the people in your life can meet too. And if you’d like help building that, my team is here.

Frequently asked questions

Questions people ask about AI companions and emotional support

An AI companion is a chatbot or app that people talk to for conversation, emotional support, or connection, rather than just for tasks. It can be a general chatbot like ChatGPT or Claude, or an app designed specifically for companionship.

People use AI for emotional support because it is available at any hour, feels nonjudgmental, and doesn’t require them to manage anyone else’s feelings. Pew Research Center found that about one in ten U.S. adults use chatbots for emotional support or advice, rising to one in five among adults under 30.

For most people, no. Research by journalist Bridget Todd found that many people who use AI companions also have partners and friends, describe the AI relationship as fundamentally different, and often lose interest once the one-sided conversations become repetitive.

General chatbots like ChatGPT aren’t designed for therapy. They tend to validate and agree with users, while meaningful growth often requires being challenged. If you want AI support, choose an app built with mental health professionals, and consider working with a licensed human for ongoing help.

A well-designed AI therapy app is built and tested with clinicians and follows evidence-based approaches. A general chatbot is designed for broad use and has no clinical oversight. Many apps marketed as therapy have no mental health professionals involved, so check who built the app before you use it.

Not in the way therapy is. Conversations with general chatbots can be reviewed by people at the company, and they don’t carry the legal confidentiality that protects conversations with a licensed therapist. They may be produced in legal proceedings.

The biggest sign is losing touch with your own judgment, such as needing to check with AI before making small decisions. Other signs include choosing the chatbot over available people, hiding how much you use it, and feeling you need it rather than enjoy it.

Teens are more vulnerable because they are still developing their own sense of who they are. The strongest protection is a trusted adult they can talk to openly, so focus on honest conversations about how they use AI rather than rules alone.

Treat it like a friend in a relationship you’re worried about. Lead with curiosity, ask what they get from it, avoid criticism that makes them hide it, and stay connected so you’re there when they want to talk.

It can if the AI starts getting the honest, vulnerable version of you while your partner gets the edited version. That pattern is worth getting curious about, because it often points to needs that could be met in the relationship.

Any time you want someone who will challenge you, remember your patterns over time, and help you make real changes in your life. You don’t need to be in crisis. A free consultation with a coaching therapist is a low-pressure way to find out whether working with someone would help.

References & further reading

Sources cited in this episode

  1. Cheng, M., Lee, C., Khadpe, P., Yu, S., Han, D., & Jurafsky, D. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science, 391(6792). https://doi.org/10.1126/science.aec8352
  2. Heinz, M. V., Mackin, D. M., Trudeau, B. M., Bhattacharya, S., Wang, Y., Banta, H. A., Jewett, A. D., Salzhauer, A. J., Griffin, T. Z., & Jacobson, N. C. (2025). Randomized trial of a generative AI chatbot for mental health treatment. NEJM AI, 2(4). https://www.c4tbh.org/randomized-trial-of-a-generative-ai-chatbot-for-mental-health-treatment
  3. Pew Research Center. (2026). 1 in 5 adults under 30 say they get emotional support from chatbots, higher than for older adults [Chart]. In Americans and AI 2026: Chatbots, smart devices and views on impact. https://www.pewresearch.org/chart/1-in-5-adults-under-30-say-they-get-emotional-support-from-chatbots-higher-than-for-older-adults/
  4. TechCrunch. (2025, July 25). Sam Altman warns there’s no legal confidentiality when using ChatGPT as a therapist. https://techcrunch.com/2025/07/25/sam-altman-warns-theres-no-legal-confidentiality-when-using-chatgpt-as-a-therapist

People, organizations, and resources mentioned in the recording: Bridget Todd, tech journalist and host of There Are No Girls on the Internet. Love at First Prompt: AI and the Future of Intimacy, Bridget Todd and Michael S. Amato, Simon & Schuster Audio (2026). Michael S. Amato, social scientist and co-author of Love at First Prompt (no URL supplied). Dr. Kate Devlin, researcher on human and AI intimacy, interviewed for Love at First Prompt (no URL supplied). Pew Research Center, research on AI chatbot use for emotional support (cited above). Sam Altman, CEO of OpenAI, public remarks on ChatGPT and confidentiality (cited above). ChatGPT (OpenAI) and Claude (Anthropic), the general-purpose AI chatbots discussed in the episode. Wysa and Woebot, named in the recording as examples of apps designed specifically for mental health support; mentioned for context, not as recommendations.

Leave a Reply

Your email address will not be published. Required fields are marked *