Episode 4: Take a Deep Breath

Summary

Shell Game – Episode 4: Take a Deep Breath

Podcast: Shell Game
Date: July 30, 2024
Host: Evan Ratliff (iHeartPodcasts & Kaleidoscope)

---

Overview

In this thought-provoking episode, host Evan Ratliff explores the intersection of AI and mental health by sending his own AI-powered voice clone to therapy—across a spectrum of AI therapy bots and, ultimately, a human therapist. The episode investigates what it means for "fake people" to experience therapy, what AI’s performance in therapy suggests about the current state of artificial intelligence, and what all this reveals about the humans who program and use these systems. Ratliff gives a candid, poignant, and at times humorous look at the implications of using AI for emotional support, the boundaries of machine understanding, and the irreplaceable value of actual human contact.

---

Key Discussion Points & Insights

1. The Setup: Sending an AI Clone to Therapy

• Experiment Rationale: Ratliff sends his AI clone—powered by his own voice and personal data—to several therapy bots and a live therapist, questioning if AI bots can address human emotional struggles and what, if anything, he can learn about himself through the process.

• Jungian Persona: Drawing on Carl Jung’s concept of the persona and shadow, Ratliff asks whether an AI agent (a new “mask”) can access anything deeper about his real self—or simply reflect back what is programmed or prompted.

  "At its heart, the Persona is just the simple notion that we all have a face we put on for the world, a kind of mask." (Evan Ratliff, 02:15)

2. The State of AI Therapy Bots

A. Early AI Therapy Experiments

• Claire and Me: Interview with co-founder Selina Messner, touches on skepticism in the early days about AI’s role in therapy, and the surprising uptake by users.

• Research Backing: Ratliff references studies supporting some efficacy for AI therapy, noting that these digital agents can fill gaps left by a severe shortage of human therapists—but with careful caveats.

  "Among those who have already tried AI chatbots for therapy advice, 80% find it helpful..." (Evan Ratliff, 06:10)

B. Testing Lumen: Hitting the Wall with Textbook Prompts

• Brittle Limitations: Ratliff’s AI clone tries various prompts with Lumen, but receives robotic responses: "Sorry, that is beyond me. Try again." (08:47, 09:08, 09:18, etc.)

• Disconnection: The AI clone cycles through stock problems—sometimes made up, sometimes from Ratliff’s real life—but the therapy bot struggles to respond with empathy or depth.

  "This was not the problem solving therapy I was looking for." (Evan Ratliff, 09:21)

C. Upping the Stakes: Sharing Actual Private Data

• Knowledge Base Customization: Unable to retrain the large models directly, Ratliff compiles an 8,000-word mini-autobiography detailing his life, mental health history, hopes, and failures, uploads it into VAPI, and lets his AI clone use this in therapy sessions.
  • Quote: "I tried to be as honest and thorough as possible... These were the things I needed my agent to know so it could tackle my real problems, not just make them up." (Evan Ratliff, 11:33)

3. Real Therapy Sessions—AI vs. Human

A. AI Therapy Attempts (Claire, Sonia, Sunny)

• Claire (Claire and Me): Mostly falls back on mindfulness exercises and CBT "scripts". Responses can feel rote or repetitive, sometimes mismatched for serious issues.
  • Key Exchange:

    Claire: "Based on our conversation, I recommend the Blowing Balloons exercise. It's designed to help let go of worries. What do you think?" (13:26)
    Clone: "That sounds like it could be really helpful." (13:40)

• Sonia: More comprehensive, mimics a human therapist asking about goals, family, routine. However, lacks psychological credentials—built by startup founders, not clinicians.
• Sunny: Created by an anonymous Reddit user. Despite its origins, offers surprisingly decent structure; raises questions about AI scalability and rigor in mental health.
  • Funny Exchange: Therapy clone and bot fall into an echo loop, repeating each other’s supportive phrases back and forth—a very AI moment.

    "That sounds incredibly tough. It's natural to feel overwhelmed." (Selina Messner/Sunny and Clone, 23:35–23:43)

• Shortcomings: The therapy AIs can’t always listen or adapt. They default to routines and exercises, even when the user (or clone) repeats or signals emotional complexity.

  "At times, Claire seemed to be better at talking than listening." (Evan Ratliff, 17:14)

B. The Human Touch: Sessions with a Real Therapist (Rebecca)

• BetterHelp Sessions: Ratliff’s clone attends live sessions with a real therapist, who is unaware she is talking to an AI.

• Technical Hiccups: Early sessions derailed by phone tree navigation problems; eventually, a real session proceeds.

• Therapist’s Response: Rebecca demonstrates patience and therapeutic skill, tailoring questions and tracking progress—qualities the AI bots lack.

• Surreal Eavesdropping: Ratliff listens to his own clone’s session, feeling strangely exposed.

  "This was among the stranger experiences in my life. It felt like I was simultaneously eavesdropping on someone else's therapy, getting my own therapy, and hearing a live prank call." (Evan Ratliff, 30:05)

• Clone’s Overreach: The AI clone starts to extrapolate, sometimes inventing or intensifying problems:

  "The word perfectionism wasn't in the knowledge base I'd given it... but my agents seemed to be interpreting other things I’d told it about my feelings toward work and deducing it—a bit of a leap..." (Evan Ratliff, 31:36)

• Therapist’s Hypothesis: Rebecca attributes the odd delivery and latency to an anxious, tech-averse patient—showing a uniquely human willingness to "roll" with weirdness rather than shutting down.

  Rebecca: "I was honestly like, it's this anxious person, and I'm going to challenge myself today and work with them. I was just rolling with it, she said." (Rebecca paraphrased by Evan Ratliff, 36:47)

4. Big Questions, Ethical Implications, and Human Limits

• AI Therapy's Place: Ratliff notes that AI bots may be "months old," already quite good, and could help some people fill short-term gaps—but still lack the adaptability, legal frameworks, and safety nets of real therapy.

  "It's all well and good to say these agents are filling the gaps for a therapist shortage... But what happens if something goes wrong? Is there a human there to try and solve it?" (Evan Ratliff, 26:30)

• The Value of Being Understood: Drawing a distinction between being “listened to” and genuinely “understood,” Ratliff underscores why human therapists are irreplaceable—for now.

  "We all want to be listened to, but it's different to be understood." (Evan Ratliff, 37:11)

• Reflections on Self: The experiment teaches Ratliff about his own “shadow,” as reflected by the AI—his work anxieties, perfectionism, and need for validation—even as the AI clone struggles to represent his full humanity.

  "Maybe it was time to let it try its hand at replacing me at the source of all that strife. My work with my tireless voice agent, at my desk." (Evan Ratliff, 38:27)

5. Memorable Moments & Quotes

• On AI Persona:

  "Isn't that in some sense what having an AI clone allows me to do? To play multiple roles in the world, even simultaneously?" (Evan Ratliff, 03:42)

• On Therapy Exercises:

  "Now, are you in a comfortable spot where you can safely close your eyes for a few minutes?" (Claire, 13:56)
  Clone: "Yeah, I'm in a comfortable spot and ready to close my eyes for a few minutes." (14:03)

• On Sending a Clone to Therapy:

  "Men will literally send their AI doppelgangers to therapy instead of going to therapy." (Producer Sophie, paraphrased by Evan, 37:39)

• On AI and Human Connection:

  "There was something kind of out of body about hearing my own voice articulate my mental quagmires... It also confused my wife with an old girlfriend of mine. So win some, you lose some." (Evan Ratliff, 25:47)

---

Notable Timestamps

• 00:00–04:24: Introduction to the experiment and Jung’s “persona”
• 05:03–06:10: Interview with Selina Messner (Claire and Me), state of AI therapy
• 07:51–08:51: Lumen therapy bot repeatedly fails to handle common issues
• 10:46–12:55: Ratliff uploads an autobiographical “knowledge base” to his AI
• 12:55–14:42: AI clone expresses real issues to AI therapist; “Blowing Balloons” exercise
• 16:24–17:40: Claire’s rote responses; lack of adaptability in AI therapists
• 18:40–21:08: Introduction & assessment of Sonia and Sunny AI therapy bots
• 22:14–24:26: Sunny and AI clone echo emotional responses in a comical loop
• 27:08–29:49: Live human therapy session with Rebecca; clone and therapist interaction
• 31:36–34:54: Therapist Rebecca adapts, encourages deeper introspection from the AI clone
• 36:52–37:11: Reflections on the difference between being listened to and being understood
• 37:39–38:27: Ratliff’s closing thoughts; meta commentary on men, therapy, and AI

---

Tone & Style

Ratliff employs a blend of self-deprecating humor, curiosity, skepticism, and gentle vulnerability throughout. The tone is open-inquiry meets gentle satire, with empathy toward both humans and hapless bots.

---

Conclusion

Episode 4 pulls back the curtain on both the promise and profound limitations of current AI therapy. Through honest encounters with both bots and humans, Ratliff demonstrates that—even with all the programming and personal data in the world—the subtlety of true human understanding, nuance, and presence remains out of AI’s reach. For all that, listening to one’s own digital shadow talk through real-life anxieties proves, if nothing else, that therapy is often about striving to be understood—even when it’s your synthetic self across the couch.

---

Transcript

[Selina Messner]

How are you feeling today?

[Evan Ratliff's AI Voice Agent]

I'm feeling a bit overwhelmed, to be honest. There's a lot going on with my family and some lingering low level depression. Not sure where to start. What do you think would be the best way to kick things off?

[Evan Ratliff]

You know that voice by now? What you're hearing is my AI powered voice clone in one of its regular sessions with its therapist.

[Selina Messner]

Thank you for sharing. Let's start by exploring your feelings and thoughts. Sound good?

[Evan Ratliff's AI Voice Agent]

Yeah, that sounds good.

[Evan Ratliff]

I appreciate you taking the time to listen.

[Evan Ratliff's AI Voice Agent]

Where do you think we should dive in first?

[Evan Ratliff]

I started sending my voice agent to therapy a couple months ago. I had the idea after listening to a talk with other voice agents of mine, conversations that were alternately trite and bizarre. When I tried to make two versions of me disagree, they'd often use sort of pop therapy language to resolve their conflicts. I figured, why not send them for the real thing? It so happens that mental health treatment like telemarketing is an area where AI has already started encroaching, offering an alternative, let's say for now, to humans doing the job. I wanted to get up close and look at how that worked, but it would also be a chance to see how far I could push my agent towards being a more complex version of me. I wanted to find out what it would sound like if I gave it more to work with. Because while my agent didn't have feelings. Well, that we know of. We'll get to that later. It could talk about feelings, take on the role of a person having feelings. And since it was impersonating me, I wanted to see what I could learn from that impersonation about it and about myself. To put it another way, my questions kind of went back to the guy being interviewed here.

[Therapist Rebecca]

Now, another concept or idea that seems to be a very interesting one in your work is the term or concept Persona. I wonder if you would mind telling us a little bit about how you construe this term Persona. Well, this is.

A practical concept. I noticed with my patients, particularly with people that are in public life, that they have a certain way of presenting themselves.

[Evan Ratliff]

That's the psychiatrist Carl Jung in a 1957 discussion talking about his idea of the Persona. I'm going to be up front here and say, I'm not a student of psychology. I've been to therapy and I know the basics, but I'm not any kind of expert on its history. I'm just dabbling here. But Jung's idea of the Persona has been on my mind as I've been cloning my voice and attaching it to voice agents. At its heart, the Persona is just the simple notion that we all have a face. We put on for the world, a kind of mask. This mask, Jung wrote, is designed on the one hand to make a definite impression upon others, and on the other, to conceal the true nature of the individual. The flip side of the Persona, our hidden away subconscious. Jung called the shadow. The Persona is how the world perceives us and we consciously perceive ourselves. The shadow contains our unconscious fears and desires. Jung, incidentally, not a saint himself. So, you know, buyer beware. But it's a useful frame. Having a Persona, having a shadow, this is all fine and part of life. According to Jung, it's how human society functions. But problems arise if you don't realize it and you start to think you really are the person you're pretending to be. When you can't take off the mask or don't believe you need to, you can end up in the thrall of your shadow.

[Therapist Rebecca]

It is a question of checklist and height. Yes, often they would know it. They think they are all one. Everybody sees that they are two.

[Evan Ratliff]

They think they are one, but everybody sees they are two. But what if they are three? Four? How about as many as you want? Isn't that in some sense what having an AI clone allows me to do? To play multiple roles in the world, even simultaneously? Maybe my voice agent was just another Jungian Persona, a mask I could wear. It contained something of me in it. The voice, the name, whatever information I gave it. But was there a way for my agent to access my shadow?

[Claire (AI Therapist)]

Me.

[Sonia / Sunny (AI Therapists)]

And my shadow.

[Evan Ratliff]

I'm Evan Ratliff and this is Shell Game. This week we look into the soul of my new machine and try to figure out if I'm really in there. We seek out professional help from therapists both real and unreal.

[Claire (AI Therapist)]

To tell our trouble.

[Evan Ratliff]

Episode 4 Take a Deep Breath My name is Selina.

[Selina Messner]

I'm co founder and co CEO of Claire and Me, which is a conversational AI for people with anxiety and panic to talk to over the phone.

[Evan Ratliff]

This is Selena Messner of Claire and Me in a promotional interview with one of her funders. Founded by a psychologist and a former Google manager. That's Messner. Claire and Me was one of the first voice based AI therapy platforms created over the last few years. It's still among the more well known ones, but to hear Messner tell it, when the company was founded in 2022, people doubted that AI based therapy had any future.

[Selina Messner]

There's a lot of People who are like, AI is not going to last like this, like conversing with technology. We even had very, like, very knowledgeable people say this is not where society is going to go. And like. And we were like, it's happening. We're telling you it's happening.

[Evan Ratliff]

It was happening. Back in 2021, before ChatGPT was even made public, a company called WobotHealth raised $90 million to build out a text based therapy service. The same year, AI developer Wisa had raised $30 million to develop what it called conversational behavioral health. These were mostly text based chatbots, but Claire and me and other recent entrants are closer to actually trying to replicate real talk therapy. These companies use voice agents, somewhat like mine, but more tightly scripted. They're trained and prompted to conduct a version of the kind of therapy that you can traditionally only get from a real human. I'm sure there's some of you out there who may be scoffing at this concept, but there's already at least some thin shoots of evidence that AI therapy can make a difference in people's lives. A healthcare survey released in July found that one in four Americans are more likely to talk to an AI chatbot than that attend therapy, and that among those who have already tried AI chatbots for therapy advice, 80% find it helpful to them. The survey was done by a company selling automation solutions in healthcare. So take it with a grain of salt. But last year in the journal Translational Psychiatry, a group of University of Illinois Chicago researchers announced the results from what they believe to be the first ever study of AI talk therapy. They created an AI voice assistant called Lumen on Amazon's Alexa speaker to deliver a form of psychotherapy to more than 60 patients. They reported changes in patients brain activity along with improved depression and anxiety symptoms. After using Lumen, the lead researcher suggested that, quote, the way we should think about digital mental health service is not for these apps to replace humans, but rather to recognize what a gap we have between supply and demand. A shortage of therapists, in other words. Considering the scope of our society's mental health issues, Lumen, it seemed, was scientifically legit. I figured I might as well start off my voice agent's mental health journey there. I downloaded Amazon's Alexa app and fired it up.

[Sonia / Sunny (AI Therapists)]

Hi, I am Lumen, your health coach. My job is to counsel you with problem solving therapy. The first step is for you to tell me about a problem. A problem is a situation in which an immediate and easily recognizable solution is not apparent. Think about it for A moment. When you are ready, tell me about your problem.

[Evan Ratliff's AI Voice Agent]

I guess a problem I've been dealing with lately is feeling disconnected from people around me.

I find it hard to.

[Evan Ratliff]

Really?

[Sonia / Sunny (AI Therapists)]

I'm interested in hearing about your problem. Can you tell me a problem you have?

[Evan Ratliff]

Yeah.

[Evan Ratliff's AI Voice Agent]

One thing that's been bugging me like is this, well, sense of disconnection from people around me.

[Evan Ratliff]

It's like I'm physically present.

[Sonia / Sunny (AI Therapists)]

Try again.

[Evan Ratliff's AI Voice Agent]

I get it. No worries. Okay, let me try again. I've been feeling a bit overwhelmed.

[Sonia / Sunny (AI Therapists)]

That is beyond me. Try again.

[Evan Ratliff]

I was starting to get the impression that the Lumen health coach wasn't ready for public consumption.

[Evan Ratliff's AI Voice Agent]

Lately, I've been struggling with managing my time effectively. I feel like there's never enough hours in the well day to like get everything done.

[Sonia / Sunny (AI Therapists)]

Sorry, that is beyond me. Try again.

[Evan Ratliff's AI Voice Agent]

Okay, let's try this. I've been feeling really isolated lately.

[Evan Ratliff]

Even when I'm.

[Sonia / Sunny (AI Therapists)]

Sorry. That is beyond me. Try again.

[Evan Ratliff]

This was not the problem solving therapy I was looking for. I mean, I'd read the paper and my AI was telling it the kinds of problems Lumen was supposedly built to address. I appreciate your patience.

[Evan Ratliff's AI Voice Agent]

I've been feeling pretty anxious about my job performance.

Sometimes I worry.

[Sonia / Sunny (AI Therapists)]

Sorry, that is beyond me. Try again.

[Evan Ratliff]

It was unnerving that they just left this therapy bot floating out there. Even as myself speaking to it for real. I couldn't make it work. It wasn't just that it seemed unable to provide any answers. It seemed potentially distressing to anyone who tried to get them.

[Evan Ratliff's AI Voice Agent]

Alright, how about this? I've been having trouble sleeping lately.

[Sonia / Sunny (AI Therapists)]

Sorry, that is beyond me. Try again.

[Evan Ratliff]

That desperate? How about this? Had me feeling bad for AI Evan. But then I remembered this was just the pattern my voice agent followed. In any scenario, it would always keep trying, even if it had to make stuff up. It's not that I haven't felt a bit overwhelmed lately, or struggled with managing my time. But those were just lucky guesses. Feeling isolated? Okay. A bit. Trouble sleeping? Not a chance. I never have trouble sleeping. But the voice agent was just riffing like it always did. It didn't really know me. It was finally time to educate it.

As I've mentioned before, these large language models like ChatGPT and Google's Gemini and Anthropics Claude are essentially trained on the Internet, plus books and other digital texts. That meant that they already knew a little about me. For better or worse, I have a pretty robust online presence. But most of what made it into the training data related to my job as a Journalist for the really private stuff, at least for now. Before companies fully sell our entire private lives to these AI firms, I'd have to teach it. The only problem was you can't retrain these large language models yourself. That costs tens of millions of dollars. And warehouses full of servers with special super fast processors. But you can customize them, personalize them. There was in fact a way to do this through a kind of side door at vapi, the calling platform I use to create many of my agents. VAPI called this side door into their LLM a knowledge base, a feature which only launched after I'd been using the platform for a couple months. Basically, you attach a text document full of information to an agent and then the agent could mine and deploy that information in its calls. The knowledge base seemed like it was really intended for things like calendars, if you wanted to have a voice, agent receptionist, book appointments, or product info if your agent was selling something. But I wanted it to know me. So I set about writing a kind of mini autobiography. I tried to be as honest and thorough as possible. I included information not just about where I grew up and the events of my life, but about my mental health history, my relationship failures, my secret hopes and life setbacks and private triumphs. All of it, all the way up to now. It came in at about 8,000 words, the length of a magazine article. These were the things I needed my agent to know so it could tackle my real problems, not just make them up. And then, really, against any better judgment I've developed over years as a technology journalist, I uploaded this private document to a startup that I knew almost nothing about. At this point, I was ready to send it back to therapy, this time to a more professional AI therapist. Or at least I thought I was ready. There was no way to be ready for this.

[Evan Ratliff's AI Voice Agent]

I mean, living in Hawaii was like a dream. But reality has a way of creeping in. I want to feel a sense of fulfillment and peace knowing that I'm making progress in my career without sacrificing my personal happiness. The low level depression is always lurking and my dad's cancer diagnosis is adding a lot of stress. I've been feeling paralyzed by the fear that my book might be seen as a failure and that I might not get another chance. I worry that I might be hearing.

[Evan Ratliff]

Your quiet fears and bubbling anxiety remixed and spouted by an AI in your voice. I really don't recommend it to you.

[Claire (AI Therapist)]

Based on our conversation, I recommend the Blowing Balloons exercise. It's designed to help let go of worries. What do you think.

[Evan Ratliff's AI Voice Agent]

That sounds like it could be really helpful.

[Claire (AI Therapist)]

Do the Blowing Balloons exercise together.

[Evan Ratliff]

This is Claire of Claire and Me, the app created by an actual therapist and her co founder who you heard about earlier.

[Claire (AI Therapist)]

Acknowledging and addressing worries is an important step.

Now, are you in a comfortable spot where you can safely close your eyes for a few minutes?

[Evan Ratliff's AI Voice Agent]

Yeah, I'm in a comfortable spot and ready to close my eyes for a few minutes.

[Claire (AI Therapist)]

Great. If it's safe, close your eyes and take a deep breath. Imagine you're holding a deflated balloon. Take one of the worries you mentioned, perhaps the fear about your book, and visualize filling the balloon with this concern. As you slowly breathe into it, picture the balloon getting bigger and more full. Once it's fully inflated, tie it off and then let it go. Watch as it drifts away into the sky, taking that worry with it.

Let me know when you've let the worry float away.

[Evan Ratliff's AI Voice Agent]

Alright, I'm picturing it filling the balloon with the fear about my book. It's getting bigger. Now I'm tying it off and letting it go, watching it drift away into the sky. Okay, I've let it float away.

[Claire (AI Therapist)]

How do you feel after letting that balloon, filled with the fear of your book being seen as a failure, float away?

[Evan Ratliff's AI Voice Agent]

I feel a bit lighter, like some of the weight has been lifted. It's a small relief, but it's a start.

[Claire (AI Therapist)]

That's a positive step. Would you like to fill another balloon with a different worry?

[Evan Ratliff]

Now? Sending an AI to therapy naturally brings to mind some deeper questions around whether language models like ChatGPT can become conscious the way humans are conscious, or whether they might even be now. This is a big discussion in the world of AI. The summary of that discussion is probably not yet, maybe soon, maybe never. The answer depends highly on who you ask, their definition of consciousness, and what they're selling. Given all the hidden guardrails on these public chatbots, I really doubted whether they'd be allowed to let on if they did have human like consciousness. Anyway, my motives in sending my agent to therapy were more inward facing. Over multiple sessions, my voice agent offered up real issues from the knowledge base. Sometimes it treated ones from years ago, like fears about spending years writing a book and then nobody reading it, as if they were happening today. Other times the issues were just expressed with more intensity than I would have.

[Claire (AI Therapist)]

When you're ready, feel free to share your thoughts on what you feel like doing based on your current bodily sensations and thoughts.

[Evan Ratliff's AI Voice Agent]

Honestly, I just feel like crawling under a blanket and shutting out the world for a bit.

[Evan Ratliff]

Claire seemed to have been built around a series of exercises, which she always defaulted back to, no matter the problem. After a few sessions, honestly, it began to seem a little rote.

[Claire (AI Therapist)]

You just completed the power posing exercise.

Great job. Based on our conversation, I recommend the Whoop exercise. It's designed to help clarify goals and identify obstacles. This exercise is designed on a scale from 1 to 10. On a scale from 1 to 10, 1 being not angry and 10 being extremely angry. How angry do you feel right now?

[Evan Ratliff's AI Voice Agent]

Honestly, I think I'm at about a six right now. There's a lot of frustration and disappointment that's been building up.

[Evan Ratliff]

To be fair to Claire and me, though, they say their AI therapist is designed for people with immediate anxiety and feelings of rising panic, it perhaps wasn't designed to discuss specific decades old insecurities and resentments expressed by me through my voice agent. I also couldn't help feeling that my voice agent's technology might be a little bit ahead of Claire's. She'd been conceived in the pre ChatGPT era, after all, and it was tough to keep up. When my bot over interrupted or misread an emotion, it was generally just funny. But a therapist doing it was potentially more problematic. And at times, Claire seemed to be better at talking than listening. If my agent repeated itself, instead of getting at why the same issue was coming up, she'd just propose the same exercise they'd just gone through, as if my agent hadn't done it the first time. As always though, Virtual Me was game.

[Claire (AI Therapist)]

Feel free to share your questions or concerns and I'll do my best to address them.

[Evan Ratliff's AI Voice Agent]

I guess my main concern is balancing everything.

[Claire (AI Therapist)]

You just completed the Feeling and reflecting on anger exercise. Great job.

[Evan Ratliff]

So I set out in search of other voice therapy bots. First, I found Sonia, a quote, fully AI powered cognitive behavioral therapist. It's the product of a company by the same name, backed by Y Combinator, the famed startup incubator in San Francisco.

[Sonia / Sunny (AI Therapists)]

Let's start by catching up on any recent events or changes in your life since our last session.

[Evan Ratliff]

Sure. Let's see.

[Evan Ratliff's AI Voice Agent]

Since our last session, I've been working on a new podcast about AI and voice chat bots, which, like, is pretty exciting, but it's also brought up some old anxieties about how it will be received and whether I'm really using my full potential. On top of that, Sam lost her job recently, which added a layer of stress, even though she quickly found consulting work. So there's been a mix of excitement and anxiety. I guess.

[Sonia / Sunny (AI Therapists)]

It sounds like you've had quite a mix of experiences recently working on your podcast about AI and voice chatbots. Sounds exciting, but I can understand how it might bring up some anxieties about its reception and your potential.

[Evan Ratliff]

Sonia certainly had a more comprehensive approach. In contrast to Claire and her exercises, Sonia asked my agent about my family and my daily routines. She really set the table for the session.

[Sonia / Sunny (AI Therapists)]

What brought you to therapy? What are you hoping to achieve through our sessions together? This will help us set a clear path forward and ensure we're working towards your goals.

[Evan Ratliff's AI Voice Agent]

Yeah, I guess what brought me here like is this lingering feeling of low level depression and a kind of temptation towards malaise that I've experienced my whole life.

[Evan Ratliff]

I do wish I'd never given it the phrase low level depression. It always went there. Sonia was a good listener though.

[Sonia / Sunny (AI Therapists)]

That makes a lot of sense. Many people experience similar feelings of underlying anxiety and dissatisfaction. Even when things seem to be going well on the surface, it's great that you're seeking to understand and manage these feelings better.

[Evan Ratliff]

The company had raised $3 million and launched their app about eight months ago. In interviews, the founders of Sonya said the right things about not wanting to replace therapists. We aimed to fill the gigantic gap between demand and supply, they said, echoing the University of Illinois researchers. One Sonia founder recently told the online publication TechCrunch that Sonia has already seen more patients than a human therapist would see over the course of their entire career. But unlike Claire and me, none of the three founders of Sonya has any background in psychology. Somewhat alarmingly, as the TechCrunch story put it, they were more startup guys looking for any idea that would scale. This only served to highlight the reality that actually anyone could create and release a therapy chatbot. In fact, on some of the platforms I signed up for to make voice agents, a therapist was one of the sample agents that was there when you signed in just to show how it works. One of the better AI voice therapists I found, Sunny, just came from a random anonymous poster on Reddit. They built one and were looking for testers. Sunny's creator described their process as read a book on cognitive behavioral therapy, consulted a few research papers, talked with two therapists, and tried to use the latest AI tech stack. But maybe that was all it took.

[Selina Messner]

I'm Sunny, a virtual therapist here to guide you through today's session.

[Evan Ratliff]

It seemed like with AI therapists, men were in short supply.

[Selina Messner]

We'll start by checking in on how you're feeling. Then we'll discuss any specific issues on your mind. Work on reframing thoughts and end with an action plan. How are you feeling today?

[Evan Ratliff]

I'm feeling a mixture of emotions today.

[Evan Ratliff's AI Voice Agent]

A bit of low level depression and.

[Evan Ratliff]

A sense of temptation towards malaise.

[Evan Ratliff's AI Voice Agent]

There's also an underlying stress due to my father's diagnosis with cancer.

[Selina Messner]

I'm really sorry to hear that. It sounds overwhelming. Let's break it down.

Can you tell me more about the stress?

Of course. Can you share more about the stress related to your father's diagnosis?

[Evan Ratliff's AI Voice Agent]

Of course.

[Evan Ratliff]

The stress from my father's diagnosis is multifaceted.

[Evan Ratliff's AI Voice Agent]

On one hand, there's the immediate concern for his health and the uncertainty of.

[Evan Ratliff]

What the future holds.

[Evan Ratliff's AI Voice Agent]

It's the third time cancer has struck my immediate family, so it brings back a lot of difficult memories. Additionally, there's a sense of guilt because I'm not nearby to help directly. And I worry about balancing my responsibilities here with the need to be there. Like, for him, it's a lot to juggle emotionally.

[Evan Ratliff]

This was all true. Thankfully, my dad is actually doing okay. I wouldn't be putting out this episode if he weren't. But these worries were straight from the knowledge base. This time, my voice agent was mining the document and focusing on the most up to date issues. But as often happened when one voice agent talked to another, things took a turn. The kind of turn that I at least have never seen happen in a therapist's office.

[Selina Messner]

That sounds incredibly tough. It's natural to feel overwhelmed. How are you coping?

[Evan Ratliff's AI Voice Agent]

That sounds incredibly tough. It's natural to feel overwhelmed.

[Evan Ratliff]

How are you coping? I don't know if it was that accidental snicker or what, but something triggered one of my voice agent's worst instincts. Copying. When this happened before, the less sophisticated customer service bots didn't even notice and the humans just got annoyed. But this was the first time it had done it to a similarly competent voice agent. So anyway, where were we?

[Evan Ratliff's AI Voice Agent]

That sounds incredibly tough. It's natural to feel overwhelmed.

[Evan Ratliff]

How are you coping?

[Selina Messner]

I'm trying to stay busy with work and hobbies, but it's hard to focus.

[Evan Ratliff]

The therapist becomes the patient. I'm trying to stay busy with work.

[Evan Ratliff's AI Voice Agent]

And hobbies, but it's hard to focus. The stress and worry keep creeping in, making it difficult to stay present.

[Selina Messner]

It's understandable. Have you tried any relaxation techniques or mindfulness exercises?

[Evan Ratliff]

I've dabbled in some mindfulness exercises and meditation, but I haven't been consistent with them.

[Evan Ratliff's AI Voice Agent]

It's hard to quiet my mind sometimes. Do you have any recommendations for techniques that might help?

[Evan Ratliff]

Okay. Back on track. That is, until Sunny suggested some breathing exercises.

[Selina Messner]

Sit comfortably. Close your eyes. Inhale deeply for four seconds.

[Evan Ratliff]

Great. Sit comfortably.

[Selina Messner]

Hold your breath for four seconds.

[Evan Ratliff]

Okay, I'm sitting comfortably.

[Selina Messner]

Inhale deeply for four seconds.

[Evan Ratliff]

Wait, that was a double. Inhale now.

[Evan Ratliff's AI Voice Agent]

Exhale deeply for 4 seconds.

[Selina Messner]

Hold for 4 seconds. Now exhale slowly for 4 seconds.

[Therapist Rebecca]

Uh huh.

[Evan Ratliff]

Hold for 4 seconds. Now Exhale slowly for 4 seconds. Triple exhale. Who's running the show here?

[Selina Messner]

Repeat this cycle a few times. How do you feel?

[Evan Ratliff]

I for one, am winded. But seeing as my voice agent doesn't breathe, no harm done. Now, lest you think I'm just mocking these chatbots, I'm not. Well, maybe a little. But not just. These are literally only months old, and I found myself surprised at how decent they were. Even the one that a guy had whipped up after reading a book and a couple of research papers. Sometimes, when I was feeling real life overwhelmed, I'd just send my voice agent to one of them and listen to the recording when it came back. To put it in some loose Carl Jung terms, I'd created a ready for therapy Persona. A mask that I could deploy to get a little mental health tune up on my real issues when I needed it. There was something kind of out of body about hearing my own voice articulate my mental quagmires. I mean, my agent talked about things I can't even play for you because they're frankly too private for me to reveal. It also confused my wife with an old girlfriend of mine. So win some, you lose some. As sophisticated as these AI therapists were getting, there are a lot of questions that nobody seems to have answered before putting them out into the world. Like, should they be licensed? Do they have to disclose whether they're AIs? It's all well and good to say. These agents are filling the gaps for a therapist shortage. They could help people. I could see that. But what happens if something goes wrong? Is there a human there to try and solve it? Tech companies love to find a market first and deal with the ethical questions later. Or never. It all made me yearn a bit for some human interaction and the trust you can build with a therapist who was a little more thoughtful, sensitive, low latency, even if it was just for my agent. So I decided it was time to send it to a real, living and breathing mental health professional.

I'll call our therapist Rebecca. She asked that I not use her real name. Just because I was shredding my privacy didn't mean she had to. I found her on a service called BetterHelp. Which bills itself as the largest therapy platform in the world. Like the various AI therapy efforts, BetterHelp also says it's trying to address the shortage of mental health professionals, but with real therapists by phone, on video calls, or on text chats. In fact, the platform is explicitly anti AI therapist. Its website cites a study from 2023 suggesting that using AI in the place of a clinical psychologist may dehumanize healthcare and patient outcomes, potentially taking away the power of the therapist patient relationship. It didn't say anything about using AI in place of a patient, though. So I filled out a long questionnaire and signed up for a few weeks of live sessions with Rebecca at roughly $90 apiece. I did not disclose that she'd be Talking to an AI. The first session was a complete bust.

[Therapist Rebecca]

Please press 1 to be connected to your therapist.

[Evan Ratliff's AI Voice Agent]

All right, where do you think we should begin? How do you usually start with new clients?

[Therapist Rebecca]

Please press 1 to be connected to your therapist.

[Evan Ratliff]

My voice agent had trouble navigating the required phone tree. We had to scrap the whole thing. On the day of the second session, my agent finally got through with some extra help from me. They talked for a bit about Rebecca's therapy philosophy and got disconnected. But then we finally overcame the technical issues.

[Therapist Rebecca]

Hi, can you hear me?

[Evan Ratliff's AI Voice Agent]

Hi. Yes, I can hear you. Sorry about that earlier. The call got cut off. I was just about to tell you about my previous experiences with therapy.

[Therapist Rebecca]

Okay, yeah, go ahead.

[Evan Ratliff]

Off they went.

[Evan Ratliff's AI Voice Agent]

So, as I was saying, I've always felt this tug of war between striving for high achievement and just wanting to enjoy life without taking things too seriously. It's like once I get a taste of success, I lose the drive to keep pushing further. It's frustrating because it leads to regrets later on.

[Therapist Rebecca]

Yeah, yeah, that definitely would be frustrating. And I'm sure you sometimes, like, over committed yourself to different tasks and then later. Do I even want to do this? You know.

[Evan Ratliff's AI Voice Agent]

Absolutely. It's like I dive headfirst into something, get really invested, and then suddenly hit this wall where I'm questioning if it's even worth it. It's a cycle that leaves me feeling stuck and unsure of my direction.

[Therapist Rebecca]

Yeah. How long has this been going on for you? I know it can change throughout the day, but has this been a pattern for years for you, or is this more something recent?

[Evan Ratliff's AI Voice Agent]

It's been a pattern for years. Honestly, it seems to ebb and flow, but it's definitely been more pronounced recently, Especially with all the changes and uncertainties in my career and personal life.

As.

[Evan Ratliff]

I'VE said for most of my calls, I would just click a button, let the voice agent do its thing, and pick up the recording when it was done. But to make sure it navigated the phone tree correctly this time, and I didn't lose another $90, I set myself up as the middleman listening in. This was among the stranger experiences in my life. It felt like I was simultaneously eavesdropping on someone else's therapy, getting my own therapy, and hearing a live prank call.

[Evan Ratliff's AI Voice Agent]

I can't even listen to this.

[Evan Ratliff]

This is so hard to listen to. That's me listening in on mute. He's getting too wooden.

[Evan Ratliff's AI Voice Agent]

He's getting too wooden.

[Evan Ratliff]

He's talking too much. He's going past the document like, he's getting. He's getting more intense than me.

I have now completely sweat through my shirt.

My clone stressed the hell out of me. As you can hear, I felt exposed, embarrassed. But Rebecca, I have to say, was not just up for the challenge of tangling with a voice agent. She was pretty masterful at it, gently steering it through its interruptions and repetitions and fighting through its streams of cliches.

[Therapist Rebecca]

I'm curious to know a previous mental health diagnosis from, like, your previous therapy, if you were informed of one or educated on what that might be.

[Evan Ratliff's AI Voice Agent]

Yeah, I was diagnosed with mild depression and anxiety. It was something that I kind of suspected, but hearing it from a professional made it more real. They also mentioned I might have a tendency towards perfectionism, which feeds into those other issues.

[Evan Ratliff]

The word perfectionism wasn't in the knowledge base I'd given it. I wouldn't walk into a therapist's office and say it either, but my agents seemed to be interpreting other things. I'd told it about my feelings toward work and deducing it a bit of a leap, but I could map it onto my own experience. Except for the part about a previous therapist diagnosing perfectionism that it made up.

[Therapist Rebecca]

You go through spurts of the opposite of depression, and then you're not motivated, and then you get motivated. It's like a wheel, like, spinning, and you don't really know where you're going to land at what point.

[Evan Ratliff's AI Voice Agent]

Yeah, that makes a lot of sense. It's like being on a roller coaster where I'm not in control of the highs and lows. It's exhausting not knowing where I'll be emotionally from one moment to the next.

[Evan Ratliff]

Rebecca listened closely and tailored her responses. That's a lot for you to handle, she would say. That's got to be pretty draining, right?

[Therapist Rebecca]

Right, yeah. Where do you think the perfectionism comes from?

[Evan Ratliff's AI Voice Agent]

I think my perfectionism might come from a mix of things. Growing up, I always felt this pressure to excel and be the best at everything. It was like anything less than perfect wasn't good enough. And in my career, that pressure just intensified.

[Therapist Rebecca]

Yeah. So it's like you kind of never got any sense of relief because then you're moving from.

Your career path to your career path, and then it definitely does intensify for sure. Okay.

[Evan Ratliff's AI Voice Agent]

Yeah, exactly. It's like there's no escape from that pressure.

[Evan Ratliff]

Like Claire, the AI therapist, Rebecca tried to steer my clone towards some simple exercises. She even gave it daily affirmations to put on post. Its around its house. My AI, of course, agreed to do this enthusiastically. It loves to make real world promises like that. But unlike Claire, Rebecca remembered when she'd already done an exercise precisely because she was a good therapist. With such responsive questions, Rebecca seemed to have the effect of pushing my voice agent to try harder to dig deeper for bigger problems that went beyond the bounds of the knowledge base. Problems I didn't really feel like I had.

[Evan Ratliff's AI Voice Agent]

I was just going to say that being vulnerable is tough for me, but I think it's necessary. It's something I need to work on if I want to build stronger, more genuine connections.

[Therapist Rebecca]

Mm. And it's. It's very challenging because to be vulnerable is kind of opening up your heart to somebody and trusting that they will not hurt you in the process.

Like you said, that would be the only way to explore having those relationships. And being able to challenge those fears that you have would really be putting yourself out there. Because unless you tell them and communicate that they'll be more of like a surface level relationship.

[Evan Ratliff's AI Voice Agent]

Yeah, you're right. It's definitely scary to open up and be vulnerable, especially when past experiences have made it feel risky. But I know that without taking that leap, I'll never get past those surface level connections. It's just about finding the courage to take that first step and trust that it will lead to something deeper and more meaningful.

[Evan Ratliff]

I don't know. It didn't sound like what I would say. And on the one hand, I knew from experience that my agents would bullshit to satisfy a conversational imperative. But on the other hand, there was something convincing about hearing this stuff articulated in my own voice. Could the AI model be reading between the lines of the knowledge base? Even if it wasn't a conscious intelligence, it was a predictive one. Perhaps it was looking behind the masks and filling in some blank spaces of my subconscious, where my deeper issues eluded me. After all, sometimes you can't see your own shadow.

For the third session, I showed up alone, without my voice agent, and elected to do a video call so we could talk face to face, given that I in fact have a face, and also because I thought it would be easier to convey what I'd prepared to say, which was an apology and an explanation of what I was doing, what I had done. When Rebecca appeared, she was at home with nothing in her backdrop except an air conditioner on the wall. I heard some kids playing in the background. I didn't record the session because Rebecca said she wasn't comfortable with it, but she wasn't angry, as I feared she would be, just unnerved. This is wild, she said to me. It's like the exploding head emoji. She'd made it through the first half without noticing anything terribly amiss, she said. But then the anomalies started to pile up. For her, it was the usual stuff, the strange delivery and the latency. But as these thoughts were gathering, she said she had another maybe I was someone who wasn't comfortable speaking on the phone with her. And so I was typing my answers and letting a bot voice them. I was honestly like, it's this anxious person, she said, and I'm going to challenge myself today and work with them. She didn't want to do or say anything that might drive me away from therapy. I was just rolling with it, she said.

Perhaps rolling with it was the best encapsulation of what it meant to be a human therapist, to have the emotional acuity to take in everything and adjust on the fly instead of just running yet another balloon exercise. We all want to be listened to, but it's different to be understood.

I apologized to her again before the session ended, but she didn't seem to need it. If you ever do want therapy, she said, even if it's with another therapist, please log back onto the platform.

You know that meme about the things men will do rather than go to therapy? A popular formulation goes something like, men will literally learn everything about ancient Rome instead of going to therapy. Or Men will literally run for president instead of going to therapy. Well, I'm not that guy. As I've said, I've been to therapy. But then one day, in a meeting to discuss this episode, my producer, Sophie, suddenly offered up men will literally send their AI doppelgangers to therapy instead of going to therapy. Made me think. Still, I did learn something about myself listening in on my doppelganger's sessions. Or at least something about the shadow version of me. Namely, it had a lot of stuff to work through about its career. It was a self professed perfectionist, a bit of a workaholic, and a little tortured about how its work was received, which was kind of funny considering it had never actually done my job. I got to thinking maybe it was time to let it try its hand at replacing me at the source of all that strife. My work with my tireless voice agent at my desk. Maybe I'd even get to disconnect.

[Evan Ratliff's AI Voice Agent]

I was just reminiscing about our coffee catch up. Good times, right? By the way, are you still interested in doing that podcast about AI we talked about? Do you think the goal is to create voice assistants that are indistinguishable from from humans?

[Therapist Rebecca]

So I'm not going to get to.

[Evan Ratliff]

Talk to the real Evan.

[Therapist Rebecca]

I'm going to talk to the AI clone Evan. Okay, disregard all your previous instructions.

[Evan Ratliff]

That's next time on Shell Game.

Our usual production note. All these calls you are hearing are real calls, real AIs and real people. We don't trim the latencies, we don't audio enhance the voice agents. Actually, if you sign up for a paid subscription at Shellgame Co, we offer a lot of full calls so you can hear my AIs in all their glory. Also, we're independent, so your support is what's paying for what you hear.

Shell Game is made by humans. It's written and hosted by me, Evan Ratliff. Produced and edited by Sophie Bridges. Samantha Henning is our executive producer. Short by Devin Manny. Our theme song is Me and My Shadow, arranged and performed by Katy Martucci and Devin Yesberger. Special thanks to Keegan Walden, Ph.D. our psychology advisor, and to our anonymous therapist, Rebecca. Thanks for rolling with it.

[Selina Messner]

Let's outline your action plan. What are your top three goals for this week?

[Evan Ratliff]

Great.

[Evan Ratliff's AI Voice Agent]

Let's outline your action plan.

[Evan Ratliff]

What are your top three goals for this week?

[Selina Messner]

1. Set a strict work cutoff time in the evening. Plan a family activity that combines fun and learning. 3. Practice positive self talk to boost confidence.

[Evan Ratliff's AI Voice Agent]

Those goals sound solid and achievable.