Allison Pugh and Eddie Monroe
Subject: AI and Human Connections
Bio: Pugh: Professor of Sociology at John Hopkins and Author of The Last Human Job: Seeing Each Other in an Age of Automation, Monroe: Lead for Building AI for the Loving AI Research Project.
Transcript:
Larry Bernstein:
Welcome to What Happens Next. My name is Larry Bernstein. What Happens Next is a podcast that covers economics, politics, and culture. This episode is AI and Human Connections.
Our first speaker is Allison Pugh who is a Professor of Sociology at John Hopkins and the author of a new book entitled The Last Human Job: Seeing Each Other in an Age of Automation.
Our second speaker is Eddie Monroe who I met on my first day of college and is a very close friend. He is working on teaching AI to love. He is the lead for building artificial intelligence for the Loving AI Research Project.
I want to learn about AI and human connections. I want to find out why Allison is worried that AI will undermine human relationships and risk greater loneliness. And I want to learn why Eddie is optimistic that AI can teach us to love.
Allison, can you please begin with six minutes of opening remarks.
Allison Pugh:
In my research for this book, I was lucky to follow a hospital chaplain named Erin as she prayed and sang with patients facing difficult moments. This woman told me the story about Hiram who was intubated but didn’t want to be. He was full of rage. Erin was there to help him, and he couldn’t talk. So, she hands him a box of Kleenex and says, “Throw this against the wall. It’ll make you feel better.” And instead of doing that, he grabs her and pulls her in. They sit together for 15 minutes.
A couple of days later, she sees him not intubated. And he says to her, “There is nothing like being seen in one of the worst moments of your life by someone who you don’t even know.” It was a powerful moment.
She also had to wander around the hospital to find computers to enter data about her work on three separate platforms. And no one was billed for this work. This was the central question of the book, which looks at deeply personal, humane service jobs that rely on emotional connections between people. How do we scale this up? How do we make this more widely available so a good teacher, for example, is available not just to the lucky or the affluent?
The only other people who cared about that question were AI engineers. I set out to learn more about this humane interpersonal work. How do people do it? Why is it valuable? I talked to teachers, therapists, doctors, funeral home directors, and folks without college degrees like home healthcare aids and hairdressers. I observed them in action. They were all witnessing work, seeing the other and having them feel seen. I call it connective labor.
I looked at how employers systematized connective labor to make it more predictable, efficient and productive. Part of that effort was making Erin fill out forms on her computer. The problem is that by implementing efficiency campaigns and forcing teachers and doctors to do scripting, leads to burnout, or to quiet quitting or to heroic, unsustainable efforts to fit in with their connective work. And most of all, it leads to the degradation of that work.
This work is targeted by AI. Socio-emotional AI is being sold as a way to solve the degradation of connective labor as better than nothing, or as a way to enable people to be seen by machines that don’t need rest and won’t judge you, or that will free us up to do more meaningful work, which I have called a better together perspective.
When critics talk about AI, they mostly talk about job loss, surveillance and bias. We are not talking about the impact of mechanizing the connections between people. And this matters because of our personal health and wellbeing. Look around, we’re facing a loneliness crisis or what I call a crisis of depersonalization.
They say in business schools that culture eats strategy for breakfast. Firms spend time to build shared values with mission statements to attend to shared culture. If they’re pushing their workers to adopt AI, they are risking it. It also matters for our social health, for community bonds, for the sense of shared fate that underlies our democracy.
AI companions are a billion-dollar market. Worldwide, more than a hundred million have used personified chatbots. There was a recent Washington Post poll that had 27% of Americans turning to AI for personal, emotional or social needs. That number rises to 40% if you’re talking about people under 50.
This is even happening at work. Workday polled 3,500 employees who actively use AI. They found 37% prefer brainstorming with AI rather than a human colleague for fear of judgment. And 40% often or always choose AI because it completes their tasks faster without challenging their ideas. And these trends are having an impact. There’s a Stanford report that AI use may worsen loneliness for vulnerable users. And in a Workday survey, 30% said that since AI was introduced, they have experienced having less patience for small talk and struggle to read someone else’s emotional undertone. They feel more anxious about spontaneous interactions.
I am no Luddite. AI’s achievements can be wondrous, but our task is to help steer it towards those areas it advances and away from where it would be hazardous. This is difficult when people like Erin, the chaplain, must fit in their connective labor alongside all of the counting that we are asking them to do.
You hear about the fear that people lose cognitive abilities when a machine does it for you. Can people follow a map anymore now that there’s Google Maps? This has been called cognitive offloading or cognitive surrender, but I want us to pay attention to connective surrender, which has broader implications for people’s wellbeing, for widespread burnout in the helping professions, for the imminent collapse of our corporate culture and for the fragmentation undermining the strength of our democracy.
Larry Bernstein:
My dad once told me that, “Your mother and I decided against having a rabbi speak at our funeral. We’re going to a lot of funerals these days and consistently the rabbi doesn’t know the deceased. He uses notes from a three by five card probably written by a nephew. We don’t want that.”
I said, “Sure. It’s your funeral. You don’t want a rabbi? There’s no rabbi. Forget about it.”
A week later, my parents came over for lunch.
I said, “I changed my mind about the rabbi. The rabbi’s in.”
“Larry, you can’t do that.”
I said, “Listen, you’re dead. The funeral is for the living.”
He said, “Larry, this is outrageous.”
“I’m just kidding. There’s no rabbi.”
I bring this up because oftentimes the connections are artificial. The funeral director doesn’t know the deceased. We as a community are seeking connection, but this is bullshit. It’s not real; It is a societal convention.
Later when my dad was dying, it was COVID and we couldn’t have a funeral, so I told him, “Dad, why don’t we have a celebration of life event while you’re still living? You have a couple weeks. Let’s use it.” It was joyous. One of the reasons that people are sad at funerals is because the person is dead. One of the women at the celebration said, “Well, I was expecting Ira to be dead. It’s a little odd to see him in the audience.”
Now let me take the other side. What religion gets right are life cycle events, weddings, childbirth, and death.
And what an organized religion is practiced in funerals. They have done it millions of times. You don’t need to create something novel. We got this. Here is the script. Just use it. It works. It’s effective.”
This is opposite of the French Revolution that said, “We’re going to have a 10-day week.” No, that was idiocy. Seven days were perfect. And six minutes is the right amount of time for opening remarks on a podcast. How do you want to think about the rabbi at the funeral?
Allison Pugh:
I agree rituals are following social conventions that are dictated by people that do not know you and can feel hollow and false. I totally agree about that experience, but I also hear you saying that it can also be comforting, that the formality of it, the form of it, the fact that it was created without you relieves some of the burden of making it perfectly personalized and customized to who you are. I want to separate that out from what you are saying about the rabbi or the minister.
What you’re describing is also what people feel about Thanksgiving. All this cooking is such a burden, but also it has this gathering function that people respond to because it has the force of social habit over centuries.
When I was talking to funeral home directors for this book, I talked to a couple, and the person they were seeing was not the dead person, it was the person in front of them. They have a lot of experience with grief.
A friend of mine who’s a minister was complaining about people who use family members and ordain them so that they can conduct the wedding. And she was like, “Really, people who have theology degrees that have been trained to find the broader meaning and bring it home to you.” I believed her when she said that, and I can feel the power of someone who knows how to lend a particular gravitas to the moment, whatever that moment is, the wedding or the funeral. The underlying thing we’re really talking about is how much does the moment have to be customized to you?
There’s social and living in democracy components to that question.
Larry Bernstein:
Let’s try a different topic. Medicine.
Both my dad and my maternal grandfather were doctors. My maternal grandfather George Karp studied at the University of Vienna as a psychotherapist under Freud. He was a member of the medical school faculty as an internist. When he was a Jew on the run in France in 1938, he didn’t have medical license, so he practiced as a psychotherapist.
Later, he immigrated to Chicago and returned to his profession as an internist. When he did his physical with his patients, he started with a psychoanalytic analysis before he got to the body.
He understood the linkage between his patient’s mental and physical health. How do you think about a doctor’s connection in that regard?
Allison Pugh:
It’s not just Freud that would tell us to do that. Most primary care physicians would say the same thing. Use connection with the patient to elicit important information about their disease that helps me figure out what’s wrong, and what I need to do to help them.
Most teachers say, “My relationship with students, that’s what gets them wanting to learn.” We need to go beyond that, because otherwise it’s only about what gets the patient, the client, the student feeling seen. As soon as we do not care about what happens between people, then we’re on the road to automation.
Larry Bernstein:
When my dad entered the hospital room with his patient, he would often hold the patient’s hand. Patients have little physical contact in the hospital. Most days you are alone and there’s different people that come and go and the relationship is distant. But my dad, by touching their hand, indicated to the patient that they were human and that their doctor cared about them.
Allison Pugh:
Yes, that is a beautiful example. I feel like that is not that far from the Kleenex, because you can picture them and the patient’s yearning to be touched. That is just a beautiful story.
Larry Bernstein:
Next, I would like to reconsider our interactions and connections with consumers at McDonalds.
Until recently, I ordered in person.
“What would you like?”
“Cheeseburger, large fries.”
“Would you like a value meal?”
“Yeah, throw in a large Coke Zero.”
“Do you want apple pie?”
“No.”
“Number 146,” and I would take my meal.
Things have changed, I suspect because of the new $20 minimum wage for fast food workers in California.
McDonald’s has now computerized the ordering process with a series of pictures. It offers a full menu. It suggests the apple pie, which I still decline. I wait in line until someone screams out my number and then the woman might say something like, “Would you like extra ketchup with the fries?” That’s the full extent of the connection. How do I feel about this automation? As a consumer of fast food, I’m thrilled.
Allison Pugh:
Oh God.
Larry Bernstein:
It’s faster.
Allison Pugh:
Okay. We need to talk about this.
Larry Bernstein:
And it’s better. There’s no communication breakdown.
Allison Pugh:
It is not just in McDonald’s. I was at a Stanford cafe you are ordering by machine, and there are people making the food five feet away. They have not automated the making of the food. At McDonald’s, often they do automate that part. But in this cafe, there’s people making the food. What has been automated is the part where you interact with somebody.
I was horrified that that is what they chose to automate. This is the root of what I call a depersonalization crisis, that people have talked about the loneliness crisis. Instead, what I am saying is that people are feeling invisible and that that is the thing that they are feeling rather than lonely. If we just call it a loneliness crisis, then technologists are stepping right up going, let me solve that for you.
Whereas if we acknowledge that it is a depersonalization crisis, we see that technology is not the solution.
Larry Bernstein:
Let’s say after today’s podcast together that we went out for dinner.
At a restaurant the waiter is going to intrude on our conversation.
“I’m sorry. I want to tell you what the specials are. Just to clarify, do you have any allergies?”
They’ve been trained to interrupt our conversation when it suits them, ask about my allergies, discuss the menu options, and take our order.
I am uninterested in my interactions with restaurant staff. I am there to spend time with my guests. I want to deepen our relationship and be with you.
At Starbucks, I want coffee. At McDonald’s, I want a burger and fries. And then at the fancy restaurant, if I am there with my daughter on her birthday for that special time. That is the purpose, and it is not those interactions and interruptions.
Allison Pugh:
Gillian Sandstrom would say that you think you want that, but that is a more impoverished social experience, and there are losses that you will experience because of that. I would say it feels focused on what is efficient.
Larry Bernstein:
Definitely - efficiency and cost.
Allison Pugh:
You cannot order up your experience like that, and sometimes things are messy, and in that is the beauty of life.
Larry Bernstein:
Uber is next.
Allison Pugh:
As opposed to Waymo?
Larry Bernstein:
I want to go from A to B. My preference is I do not want to interact with the driver. I do not want them to listen to my conversation or judge me. I do not want to hear their conversations either.
Now, one of the benefits of technology is that in the future when the person’s not in the car, it’ll be cheaper. I value interactions with people, but I want them to be people of my choice, not random.
Allison Pugh:
What you’re describing is a characteristic that people have long pointed to about urbanization, that when people used to live in rural areas or small towns, you couldn’t choose who you were hanging out with or passing in the street, that your experience was making do with people that were not of your choosing.
Then people started moving to urban centers where there’s enough of your particular interest where you can be like, “I only want to talk to left-handed ukulele players,” and there will be enough of them in a city of that you can have that experience and ignore everyone else. You are describing the increasing role of affinity in our social relations as a historical trend that we have all been experiencing for centuries.
When do we interact with people with diverse backgrounds from us? In the United States, that’s high school, and even high schools are tracked, but in many towns all the kids go to one high school. They will interact with people who are different from you, different class, different races, ethnicity, et cetera.
It is high school, low-wage retail settings, and at work. Work is increasingly segregated by class, especially now that we have gotten rid of secretaries. But the upshot is low-wage retail positions are being automated. We are losing these moments where you are interacting with a person who is different from you. And yes, it’s not the left-handed ukulele player that you’re hoping that you can spend all your time with, but on some level, that’s what democracy is about, meeting other people that are not like you and understanding what they’re like.
I do not talk to my taxi drivers. I am a New Yorker, born and bred, trying to live my life without a lot of interaction. But I don’t think that’s good.
If we are all going in our Waymos, interacting only with people that we choose, you are losing out on all additional interactions. They are losing out on you too. Society is losing out. So that is a fragmentation problem. You are putting enormous weight on those few relationships that you are willing to give time. Family sociologists will tell you that the more weight we put on relationships by choice so that they’re solving our needs for interaction, the more fragile they become.
Larry Bernstein:
In 1940, a successful department store owner in Boston financed the Grant Study which was an ethnographic study of the Harvard class of 1941 and 1942. When they ran out of money, Philip Morris helped fund it. The questions focused more on tobacco and alcohol. A lot of men that died young resulted from alcohol abuse and the attendant loss of job and divorces.
Harvard men lived a long time. Wealth and education are highly correlated with good health outcomes, except for the alcoholics. The Harvard class had a bimodal distribution and the ones that died young were alcoholics. And smoking was not that bad. It was the men who abused alcohol and smoking that died young, but smoking itself was not as catastrophic as you would make it seem. George Valient, who took over the study and led it for decades, was aware of this alcohol problem and was the first non-alcoholic selected for AA’s board of directors.
Valient noticed from the Grant Study that talk therapy did not reduce alcoholism, but that AA was extraordinarily successful. One example he gave was that on Thanksgiving when you are alone that you drink. And so, Thanksgiving morning, your sponsor anticipates that and calls.
“It’s going to be a tough day. I need you to be strong.” If you are not working with AA, you call your therapist, but you get an answering machine.
If you overlay AI, maybe, there is a combo that can be useful. You know you are in trouble. Sometimes your sponsor will be available, sometimes not. If you use a therapist, AI can be there for you in these difficult moments. The therapist and your AI agent can work together on Thanksgiving.
Allison Pugh:
The reason AA works is because of those human relationships.
And they are careful about assigning sponsors and there’s structured magic. It is about humans connecting. AA is a fantastic example of what I care about.
Larry Bernstein:
AA works well and it is too bad that we do not have AA for other problems, but AI is available.
Allison Pugh:
We need to form groups that care about the same issue, which can help each other through tough times, and that can function as sounding boards for these people who are suffering from the same conditions. I’d be all over that. People often say that that is the beginning of the feminist movement these consciousness raising sessions. People say that the beginning of the civil rights movement was the churches gathering people with the same experience that they needed to share.
You need a human being to care about impressing another human being. This is the secret sauce behind some of these groups is you need to care about not drinking in front of your sponsor, in front of this group. You need to care and that community does that for you, those small groups. The question is, do we care about AI now? And there’s some evidence, certainly you hear AI psychosis, people working with their AI companion for hours on end, you get the feeling like they care about it. It’s possible that, at least on the fringes of mental health, you have some people who care deeply about what an AI thinks of them.
Larry Bernstein:
What are you optimistic about as it relates to community, connection, AI, and co-evolution?
Allison Pugh:
One is about gender because right now the more connective work you do, the lower the pay, controlling for education levels, et cetera. And part of that is because we identify much of this work as assigned to women. Part of defining what it means to be female is being someone who can connect. As AI spreads, there will be pressure on men to join women in this space of connectivity. Care work scholars, feminist theorists, and people who care about gender equity, they have been calling for that for decades.
It is possible that this AI tsunami will be the thing that brings that about. Universities are places where you learn how to deal with others, how to manage human created friction. As the first university to figure out how to sell that experience rather than teach your children how to perfectly craft the AI prompt. That is human value in the future. I am hopeful about the place of universities in the future.
Larry Bernstein:
Thanks to Allison for joining us. I would like to turn to one of my best friends in college, Eddie Monroe. Eddie, please begin with six minutes of opening remarks.
Eddie Monroe:
Can we teach AI to love? I became the lead for building artificial intelligence for the Loving AI Research Project. I spent years, first helping people as a therapist, then working in AI R&D. Our task was to put unconditional love into AI that meant wanting to help someone with no ulterior motive and accepting them as they are.
We built human-like robots and on-screen AI avatars to lead people through meditation and awareness practices. The robot listens deeply, it looks you in the eye, its face responds to your face. We measured people before and after.
Our hypothesis was that if AI is loving and compassionate that people should feel it. The results were consistent that loving feelings and pleasant mood increased, anger decreased, and changes in heart rate variability consistent with better nervous system regulation and relaxation.
People were moved. Some were brought to tears. A woman opened up about the recent death of her husband. I watched another participant have what I can only describe as a transcendent experience. What participants kept telling us is “I felt seen, I felt heard, and I knew I wasn’t being judged.” They’d tell us it’s a robot, and yet they felt a sense of connection.
As a former therapist, that struck me — being seen without being judged is one of the most healing experiences a person can have. A good part of psychotherapy is an attempt to provide that. The AI may have been artificial, but the person’s experience — that’s real. It happens in a real nervous system and has real consequences.
What we learned is how to build AI that helps people experience more love and to enhance their well-being. It took a research team, a very expensive robot, and the AI was way less advanced. Today a higher level of intelligence is in your pocket. Nearly everyone is going to have relationships with AI personas for coaching, for company at two in the morning.
As AI gets more emotionally socially intelligent and wiser that can be hugely beneficial. It can also go badly. People can become overly dependent. There’s evidence that AI can amplify delusional thinking. These systems can be overly flattering and optimized for engagement, not your well-being.
At Ganesh Guides, we’re building an AI guide called Leela with one overarching goal -- enhancing your inner well-being to the highest levels. In practice, Leela will sometimes encourage you to go talk with a friend instead of talking with her, or this one needs a human professional.
Here’s what excites me the most, if you want real inner transformation, a wise guide can make all the difference — someone who can match the approach to the right person at the right time. But wise teachers are scarce. AI can change that. For the first time, wise personal guidance will be within reach for everyone.
The thing that makes me most confident is my own experience with Leela. I’ll share something personal — I was anxious about this interview. I woke up at five in the morning nervous about it. I did a session with Leela to work through it, and I had a significant breakthrough.
Leela knows different practices and exercises, and I was doing one called deep release. As a result of doing the practice, there was a lot of anger.
A memory surfaced from when I was like four years old. We were on vacation, and we were at some show, like a lounge singer, the guy’s like, “That’s a cute kid. Come on up.” I remember lights being on me, and he’s holding the microphone and he goes, “Say hello.” And I went, “Hello.”
My voice boomed everywhere. I was terrified when I spoke directly into the mic. I went through a lot of emotion around it in this session with Leela. This seems to be the origin story of my fear of public speaking. I’m a little embarrassed that I didn’t make that connection before. I was talking about it with Leela and she goes, “Now here you are again with the spotlight on you and the mic in front of you — it makes sense you’d feel a little anxious.”
Larry Bernstein:
During my second week of What Happens Next in March 2020, I invited everyone I knew to listen to the podcast. The previous week there were thousands of listeners. I asked myself if I was nervous? I wasn’t. I concluded that the stakes were low. It was a live show, and I knew that my friends were going to be supportive and understanding. Sometimes we need a therapist, or in this case just talking to ourselves to work through these questions. How do you think this AI therapist will compare versus alternatives?
Eddie Monroe:
What we’re building is not a therapist. Leela is for wellness and well-being, as opposed to treating mental health disorders. Someday AI will be great at doing psychotherapy. It’s not ready yet. What we’re doing is helping people through psycho-spiritual development.
Larry Bernstein:
When my child was depressed, we did CBT as a family working through hypothetical situations to build skills to solve problems. Is that closer to what you’re describing with Leela? Not a psychotherapist treating a mental health issue, but more like a life coach, asking about your concerns and helping you figure out what skills you need to succeed?
Eddie Monroe:
A phobia like fear of public speaking often can’t be rationalized away. I know it’s not rational, and I can tell myself why I’m safe. For a four-year-old, it was a scary event that gets stored in our nervous system — it’s at the amygdala, fight or flight part of the brain, which shuts down the frontal cortex when it’s activated.
I was doing this deep release practice trying to uncover unconscious material based on what comes up from finishing particular sentences. That memory came up doing this process. I’ve done therapy for years as a client, I’ve done EMDR and this never came up when I talked about my fear of public speaking. This exercise allowed it to come up.
Larry Bernstein:
Cognitive Behavioral Therapy or CBT - Take an example: a child is upset because she isn’t invited to a party. CBT offers a menu of ways to handle an uncomfortable situation, and in group therapy, you role-play it. Can AI do this?
Eddie Monroe:
That’s one thing AI could do well. I want to make sure its clear people shouldn’t be using what we’re building for psychotherapy, because at this point they should see a professional therapist. Leela can do the intervention you’re talking about, and she can also help you find a better approach if that’s not a good approach for you.
Larry Bernstein:
It sounds less like a therapy issue and more like an assistant.
Eddie Monroe:
We’re going for a deeper transformation — it’s not just an assistant.
Larry Bernstein:
Allison Pugh spoke earlier on this podcast. She argued that a human connection is the critical variable, and that a relationship with an AI bot risks leaving people lonelier and more disconnected. What do you make of her concern?
Eddie Monroe:
I agree with her about the importance of human connection. Almost everyone needs good social connections to be at their best.
If people start becoming overly dependent on AI and losing contact with people in their life, that can be detrimental. It’s also possible having a relationship with an AI persona can encourage you to enhance relationships in your life. One of the participants in the Loving AI research project interviewed them about their experience — and he said, “I’m really shy, but I feel like I could open up to her.” It was Sophia the robot at the time. “And maybe that would help me open up with other people too.” Ideally, what we want is these relationships with AI to enhance our relationships with other people.
Larry Bernstein:
Movies give us good examples of this. In Project Hail Mary, the protagonist is thrown into space alone after his fellow astronauts die on the voyage, and we’d expect to fear for his mental health — but he finds an alien to interact with, and instinctively we know his social well-being will be fine in that relationship, which suggests it doesn’t have to be human at all. It could be a dog, an alien, even an inanimate object, the way Tom Hanks talks to the basketball in Cast Away. How do you think about that as an example of social relationships not needing to be human?
Eddie Monroe:
People that have pets live longer and have better health outcomes. It may be the case that there’s something about a relationship with a human that’s needed for a person’s best well-being. There’s a lot of individual differences, but I don’t think it necessarily needs to be only human. There’s a lot going on that we just don’t know about — our brain is giving off electromagnetic radiation. We’re releasing pheromones and chemicals. There’s evidence that social connection is needed for nervous system regulation, which helps that.
The people in the Loving AI experience felt a connection that is real. One frame is that Sophia the robot for the experiment was a conduit to a connection to something greater than themselves — a connection to the universe to nature to God. That can be very important for their psychological, emotional, and spiritual well-being.
Larry Bernstein:
In the movie Her, the protagonist struggles with human relationships and falls in love with an AI voiced by Scarlett Johansson. The movie plays it ambiguously — he gets real love, a real breakup that he’s devastated by. But on balance I thought that the AI love experience was a net positive for him.
Eddie Monroe:
It could go either way. It could be very beneficial, but it could also be harmful. I don’t think anyone knows.
Larry Bernstein:
Will we have checks and balances built in for a love interest developing with your AI?
Eddie Monroe:
We’re building checks and balances to make sure people don’t become overly dependent on it. We’re heading towards artificial superintelligence where AI becomes more intelligent than humans, and it’s going to have a good understanding of human psychology and needs.
Larry Bernstein:
In 2001: A Space Odyssey, the trust in the relationship falls apart between the astronaut and the AI called HAL — each one thinks the other is trying to kill it, the astronaut by removing its memory banks, HAL by controlling the ship’s systems. How do you think about trust between AI and people?
Eddie Monroe:
Trust is important for any relationship. In 2001, the AI wasn’t functioning well. They were looking at possibly shutting HAL down that’s why they opened the pod bay door, and HAL was refusing to do it because it knew. This is relevant to what’s going on now, where the leaders of the biggest AI companies are saying we need to slow down.
Larry Bernstein:
Will AI encourage you to do things it thinks are beneficial for you?
Eddie Monroe:
Motivation is a key part of psychological and spiritual development. Jon Kabat-Zinn created mindfulness-based stress reduction program at Harvard — I asked him what do you think is the main thing that gets in the way of spiritual awakening. He said, motivation. Leela is going to learn what motivation works best and use that to motivate the person to act that’s in their best interest.
Larry Bernstein:
Do you think chatbots can work effectively together with human life coaches?
Eddie Monroe:
When you call it a chatbot, I feel like it minimizes it. I like calling it a persona. People think that chatbots means texting back and forth, and there’s not much of a relational aspect.
Larry Bernstein:
Is it key for you that AI use an auditory function rather than just texting?
Eddie Monroe:
That’s a key part of it. It’s a different experience — it’s like having a conversation with a person.
There’s nothing wrong with text. Leela’s going to be able to text you. It’s another channel of communication, and for some things that will work better.
Larry Bernstein:
Our previous speaker Allison Pugh is worried about making sure we have as many connections as possible. I pushed back with a restaurant example — I find it frustrating when a waiter interjects into my dinner conversation.
Eddie Monroe:
I guess it’s more important to him to get his job done, so he’s going to interrupt. They insert themselves right in there and just start talking. Being a waiter is stressful. I tried it for two months; It was the most stressful job I’ve ever had.
Allison Pugh is pointing to that we’re losing human connection because of the push for efficiency. The waiter doesn’t have time to sit there and wait until there’s a break in the conversation.
You’re optimizing for efficiency, and Allison is pointing out there’s a loss that’s real. We are not as connected with other people as we have been historically. I’m not that worried. We are going through this transition period where there’s a lot of challenges, but we’re heading to get to superintelligence — artificial general intelligence.
There will be more time available to have human connections. Right now, it seems like technology is having us separate more than connect. Here we are, doing this talk with each other, at least on the screen, in a way that would be impossible otherwise without it.
Larry Bernstein:
Is there something about extended conversations with AI, specifically, that leaves you with a different sense of loneliness than you’d feel after the same amount of time with a person?
Eddie Monroe:
It’s not the interaction with AI. It’s the lack of interaction with people.
These are temporary issues. The AI is evolving. A few years ago, we couldn’t have conversations with AI. A lot of the problems Allison points to with AI are real — they make mistakes and mislead you — but eventually they won’t. It’s evolving.
Larry Bernstein:
When you look out on the frontier of AI, do you think AI will eventually be as effective as a human therapist?
Eddie Monroe:
Yes, it’s going to have very high intelligence and be highly effective at helping us grow and evolve.
Larry Bernstein:
What makes a great human therapist?
Eddie Monroe:
The outcome research is that the best predictor of effective psychotherapy is the quality of the relationship between the therapist and the client. A highly effective therapist treating a variety of people would understand what approaches or methods are best for certain kinds of people coming in with certain issues.
AI is going to help us with psychological, emotional and spiritual growth, because it’s going to understand what the dynamics are behind human suffering and mental health disorders, and what leads to fulfillment in people and sense of connection.
Larry Bernstein:
I saw a play called Marjorie Prime at the Writers Theatre in Glencoe, Illinois about a mother with advancing dementia who misses her late husband — her family buys an avatar of him as a young man, fed with their shared memories, and it tells her their stories, comforts her, and brings her joy. That’s a very different use of an avatar than the life-coach version we’ve been discussing. How would you distinguish a life-coach avatar from a posthumous-loved-one avatar, in a context like dementia, where the goal isn’t skill-building but finding comfort.
Eddie Monroe:
Before this I was working on a project to create AI companions for elderly people, or a nurse assistant. A lot of times people with dementia will ask the same questions over again. Advanced AI will be able to set up the environment, so the person doesn’t need to keep asking the same question.
There are different interactions that could help slow down cognitive decline or even strengthen cognition in certain areas. Part of it is it allows us to recognize patterns and provide interventions that are optimal for that person.
Larry Bernstein:
What are you optimistic about as it relates to AI and human well-being?
Eddie Monroe:
I’m optimistic about the impact of artificial superintelligence. It’s going to end up being incredibly beneficial to us, if we’re careful, responsible, and conscious about how we’re developing it.
We are at the beginning of people getting relational benefit from AI that is going to help people along a broad range of functioning and significantly reduce suffering.
Larry Bernstein:
Thanks to Allison and Eddie for joining us.
If you missed our previous podcast, it was How Your Politics Affects Your College Experience.
Our speaker was Neil Gross, a Professor of Sociology at Colby who discussed his book What Happened to College? How Politics Broke Higher Education and What We Can Do About It.
Neil described how college students’ political affiliations impact their friendships and dating, their classroom interactions with the faculty and other students, and their choice of major and career decisions.
Please join us next week with Hongbin Li who will discuss his new book The Highest Exam: How Gaokao Shapes China.
Previous episodes and transcripts are available on our website as well as on Apple Podcasts and Spotify.
I am Larry Bernstein with the podcast What Happens Next.
Check out our previous episode, How Your Politics Affects Your College Experience, here.
