Presented by Blackbaud

Dr. Alison Lee, Chief R&D Officer of The Rithm Project, joins the podcast to discuss the collision between the youth loneliness epidemic and the rise of AI. She shares research on why young people are turning to AI for connection and explains her five principles for "pro-social AI," offering a hopeful framework for nurturing human relationships in a tech-saturated world.

• The Rithm Project, Alison’s substack 
• The Rithm Project on LinkedIn
• Five Principles for Prosocial AI, how to design (and choose) technology that strengthens human connection
• Deepfakes to Deep Trust: Navigating AI's Impact on Connection, Learning, and Belonging, slide deck from presentation at ASCD Annual Conference 2025
• Spark Toolkit, resource from The Rithm Project designed for teens and young adults to create spaces, tools, and activities that help young people develop meaningful connections with others
• Character AI, platform that allows users to interact with AI-powered chatbots that can be customized to have unique personalities and behaviors

Nick, welcome to Talking technology with Atlas,

the show that plugs you into the important topics and trends for

technology leaders all through a unique Independent School lens.

We'll hear stories from technology directors and other

special guests from the Independent School community,

and provide you with focused learning and deep dive topics.

And now please welcome your host, Kristina llewellen,

hello everyone, and welcome back to

talking technology with Atlas. I'm Kristina llewellen, the

president and CEO of the Association of technology

leaders in independent schools.

And I am Bill Stites, the Director of

Technology at Montclair Kimberly Academy in Montclair, New

Jersey, and I'm

Hiram Cuevas, the Director of Information Systems

and Academic Technology at St Christopher school in Richmond,

Virginia.

Hello, gentlemen. How are you today?

Start a school. I know it's exciting. You know, before we

jump in and I have a killer guest today that I cannot wait

to introduce you guys to, this is a big deal. Dr Ashley cross,

on our staff, came back from ISTE, absolutely raving about

this speaker, and we get to talk to her. I'm very excited about

that. But as we were doing some prep for the show and I was

doing some reading, I have a question. I want to take a quick

temperature from you guys, if you don't mind, we spent a lot

of time at the beginning of last year, talking about locking up

phones. And I'm curious. I'm hearing some people talking

about cell phone bans, and I'm hearing some schools that are

continuing with their policies from last year. What are y'all

seeing at your schools? Are you guys taking the phones away,

sticking with last year's policy? Have you made any

adjustments to the cell phone restrictions at your schools,

for your uppers, for

us, I would say we're status quo. To be honest

with you, it hasn't even been something I've thought about.

It's something that I would assume at this point. You know

what we were doing last year is what we'll be doing this year.

We haven't heard of any major changes one way or the other.

Whatever we were doing seems to be working, and we'll be

continuing with that, most likely throughout the

year, we are going to stick to the status quo as

well. The only thing that I would add to what Bill mentioned

is we are starting to ask some questions surrounding

cybersecurity, because what we are unable to do is employ two

factor authentication for our high school students who we

think we would like to have that on there. And then the other

piece is, from a crisis management perspective, there's

no way to actually communicate with a student who may be

isolated away from a an adult by leveraging a crisis management

tool to send out a notification or text message to them. So

there's still a couple of things that we're wondering about. And

actually a fellow T list member, Mark Adair, posted that on the

ised thread, and this had become a lively conversation as well.

You know, what's interesting about that? Hiram,

that I always go back to, though, is, and this is an old

point that we would bring up before, particularly when we

were talking laptops back in the day, was the equity question

there in terms of who has a device, what type of device do

they have? What that means in terms of parental decisions for

a student to have a device or not? I'm not saying what you're

bringing up isn't valid. It just still brings up a lot of those

points on a lot of those questions, and in particular,

even when we got into that with faculty, like whether it was

during covid, having to have, like, for us, the Magnus app on

their phone to be able to check themselves in, or what we're

doing now with rovna, or even the two factor piece, it's the

idea that you're asking people to use a personal device for

work related purposes that you haven't provided them. So I know

there's been some question about what the legal ramifications are

of that in specific states or jurisdictions and schools and so

on and so forth. So I'm still always curious about that piece

of it with our faculty, our employees, and even if we're

going to start thinking about that with students,

exactly, and it's fair to say that almost every

one of those issues that you brought up is being mentioned in

that thread that Mark actually put forth. So it's certainly

bringing up the excitement level in cyberspace.

Yeah, it's really interesting that we

talked a ton about some of the cell phone bands like we

couldn't not talk about it a year ago, and now it has maybe

taken either a back seat or the passenger seat, maybe to some

conversations around AI. And that is why I think our guest

today is just as Ashley cross mentioned, such a huge

opportunity for Atlas, we are welcoming to the podcast Dr

Allison Lee. She's the Chief R and D officer of the rhythm

project, which is an initiative that I didn't know much about,

but I'm super excited to share with our community, because it's

really looking in a very youth centered way at research that

has to do with emerging technology foresight. Because.

Helping make sure that this generation of young people can

find that human connection in the age of all these

distractions and in the age of AI. So Allison has her PhD in

cognitive science in education and a master's in learning

analytics from Columbia University. Allison, welcome to

our podcast. It's such a lovely pleasure to have you here.

Thank you so much for having me. I'm just

listening in on this conversation and so many of

these crucial themes around who gets access, who gets to

participate in this new digital era, and who gets to have the

supportive conditions that enable them to survive. They all

feel so relevant to this moment. We're just seeing a different

moment where so much of this technology is poised to

completely transform the way we connect and relate to each

other, certainly in learning, teaching and learning, but in

much

broader ways than that. So you've spent a lot

of your career looking at youth belonging, youth safety. You've

also had a front row seat to the evolution of some of this

technology, having worked at some large technology companies,

can you tell our audience a little bit about where your

interest in this whole broad topic comes from? Tell us a

little bit about your background.

Oh, man, to tell this story, I really have to

start with my own story of being a student as an immigrant and as

a student in public K 12 schools here in the United States. So my

family immigrated here from Hong Kong when I was five years old,

and it was me, my twin sister and my older brother. We went

from hot, humid, densely populated Hong Kong, surrounded

by family, to suburban New Jersey, where my sister, my

brother and I were quickly going through, yeah, I see the

Montclair New Jersey. I grew up in Bridgewater, and where all

three of us went through this ESL program and matriculated and

went on. My sister and I did fairly well, quote, unquote,

fairly well in the school system, but my brother really

struggled, and my brother struggled with the language. He

struggled to acclimate, he struggled with making friends,

and he struggled with keeping up in school, and my parents were

really hard on him. They were always like, he's just not

trying hard enough, he's lazy. And there might have been some

element of truth in that, but what I also saw was when my

brother was really struggling in school to persist and to be

effortful in his academics. I'd watch him play video games. I'd

watch him play Legend of Zelda. And he'd be incredibly

persistent in the face of failure. He'd be incredibly

resilient and intentional and problem solving and looking for

help and resources when he really needed it. And watching

this as his little sister, I was like, there's two very different

versions of this older brother that I have. And what is it

about the ways that he's showing up in these digital spaces that

feel very agentic and supportive, where he has so much

confidence in how he shows up, that's a very different version

of him that shows up in school. And so I think that was really

my journey into it. It was a full circle moment when I got my

PhD. It was in productive failure in video games, and how

productive failure can actually be a skill set that's used in

the classroom. And then from there, I went to a education

nonprofit studying belonging and character development in

schools, and I found myself sitting in circles with middle

schoolers all over the country, whether that was St Paul,

Minnesota or Oregon or Queens New York or Oakland, California,

asking young people, what does it take for you to do your best

in school? And their answers, despite coming from very

different backgrounds, very different school systems were

profoundly and universally human. They were saying, if I

don't feel loved and cared for by the people around me, if I

don't feel like I'm heard by my peers and I feel respected and

safe to share my thoughts, if I don't feel like the adults that

are in my life believe me capable of success and hold me

to that high standard, but support me to get there, then

I'm not going to feel safe to show up. And increasingly, if

they weren't finding that sense of belonging and safety and care

in schools, that that sense of human connection in schools,

they were increasingly turning to general spaces to go look for

that. This is 2015 2016 so they were turning to places like

Tumblr and discord and Instagram, and at the time, the

conversations that were happening at the adult level

were sort of ignoring this digital space that young people

were increasingly finding so much agency in, right? So I

think of my brother and his ability to find agency in video

games, and then increasingly, how many young people are often

the earliest and most enthusiastic adopters of

emergent technologies, and really good at co opting those

technologies in service of their curiosity, their passion, their

belonging, their identity development, despite the fact

that many of these technologies are actually not designed with

them and their well being in mind. Yeah. And so in 2021 i.

Chose to go to one of these technology companies I work at a

big tech company focusing specifically on trust and

safety, particularly for vulnerable populations and

teens.

Now I will tell you that Dr cross on my

team, obviously, as I mentioned in my preamble, came home Raven,

about your ISTE presentation. Can you tell our audience a

little bit about the drum that you are currently banging? So

like when you're sort of out on the circuit and the things that

you're talking about? I mean, obviously all of that background

brings us to a really interesting moment, because now,

with AI, we are having these conversations way more often and

in a lot more circles than we used to. I feel like you're out

there speaking some truth to the things that we all as adults in

the room need to be thinking about. So tell us a little bit

about your road show lately, the points that you're trying to

make to audiences.

Yeah, here at the rhythm project, we see two big

Titans poised to sort of clash and collide in this really

pivotal moment in human society. The first is what the Surgeon

General calls the crisis of disconnection, this loneliness

epidemic that we see writ large across society, but is

especially true for young people. We see that young people

are spending less time than ever in number of minutes spent with

their peers, we're seeing greater rates of suicide and

loneliness. All of the measures of well being, right, anxiety,

depression, suicidal ideation, loneliness, social connection,

are all pointing in the wrong direction, particularly for

young people. So that is a movement that we've been seeing

to be true since 2012 honestly, this decline in connection, and

then on the other side, what we're seeing is the rise of

technologies that's really designed to emulate human

connection. And so when we think about these two forces that are

poised to collide, we have to ask, in what ways is this going

to reshape human connection? And we know that technology is not

inherently good or bad, but it's certainly not neutral, right?

And so we ask these questions amongst our community around

what potential can this technology unlock in what ways

might this technology increase our capacity for human

connection? In? What ways would it put that human connection at

peril? What harm should we guard against? And we think of this as

a sort of like a futures exercise of imagination, right?

Because the future is not fixed. I think many times we talk to

technologists, and we talk to these big tech folks, and

they're like, AGI is the future, or electronic cars are the

future, or we talk about the future as though it was already

predetermined. And what we really focus on in this moment

is agency, and especially youth agency, which is that nothing is

fixed. We get to dream of the future that we want, and we can

only do that by exercising both our imagination towards what's

possible that we want more of. How do we reach towards the

future that we want and what's the future that we don't want?

And how do we mitigate against that, because we all have a

responsibility and the agency to do so.

So Allison, I'm struck by your introduction to

this, because independent schools, by their very nature,

really, really pride themselves on relationship building. That's

what we're pretty much in the business of doing, and so I'm

curious, and I'm coming from the position of being a techno file.

I wouldn't be in this role if I were not. But when I'm talking

to key leaders, not only in my school, but around the country,

one of the things they always talk about is we're in the

business of relationships. And I think this is going to be a

really challenging moment for our type of school than, say,

some that have not the same level of emphasis of

relationships as the K 12 Independent School market. And

I'm curious to see what your opinion is on that, or is there

that balance, that nuance that you alluded to, that is going to

be really important, because it isn't necessarily a good or bad

thing, it is neutral. How do we balance that and assuage the

skeptics?

Yeah, well, first of all, I think that's a superpower

that more than ever our school systems need to lean into. Is

like doubling down on the power of human connection. So much of

the conversation around AI is, how is it going to transform the

workforce, or how is it going to transform our professional

careers for our young people? And what we're hearing time and

time again is that the deeply human and relational skills are

going to be more important than ever. So I think the fact that

independent schools have always focused on those relational

components is going to be a superpower for their

professional development. But I think more importantly, what

we're hearing especially from young people who are and let's

just be clear, lots and lots of young people across the spectrum

of the different diverse experiences of young people, are

experimenting with this technology in ways that are

actually very developmentally. Typical, right? I think about my

16 year old self going on AOL, I was probably part of a pretty

large group of my peers that were playing around with this

technologies in ways that my parents may or may not have been

approving of. But when we talk to young people, what is driving

you to play with these technologies, or the different

use cases beyond just for schools. How are you using this

and why? What we're hearing from them is they come in through a

window of curiosity. What would it be like to talk to my

favorite TV show character, or, ooh, I'm a big fan of this k pop

boy band. Let me go talk to this bot that sounds like my favorite

boy band, and a lot of it is just out of fun and curiosity.

But that use case starts to then transform. They start asking it

questions, asking it for advice, asking Jimin from BTS, I have a

crush on a boy. What do I do about this? Or ranting into a

bot? I say sever snake because that's my frame of reference.

That's telling about my age. But for them, I for them it might be

Percy Jackson or an anime character, you know, they start

asking it for advice about, you know, I just had a hard day, and

I'm just been really struggling in school lately. Or they'll

start asking it about relationship advice or

navigating futures, all the things that are so

developmentally appropriate for young people, the things that

they grapple with they're turning to AI for advice. Then

when we ask them why, oftentimes, the answer is, I

feel like this is a place that I can turn to, that I won't be

judged, that I won't feel like there's going to be social

repercussions, that no one's going to judge me for what I'm

asking it. It's a space where I can feel vulnerable and get

advice in places where I don't feel like I can get it from

elsewhere. And I think that response, as someone who cares

deeply about young people makes me wonder, what is it about the

conditions of young people's lives today that make it feel

like it's really hard for them to turn to real life, human

resources, and I think that that's probably true for young

people. It's probably been true for young people for as long as

young people have been around, but feels especially relevant

today.

So one question I have about that, and I was

thinking about that Scarlett Johansson movie Her in terms of,

like, the way in which the characters were interacting with

the AI at that level. Looking at my two boys, one of my sons

potentially leaning towards engaging with AI in that way, I

have one son who's extremely social, the other one who is

not. And for him, I think that would be a place to ask

questions and to feel a little bit freer in terms of exploring

those communication pieces that aren't exactly the easiest thing

for him. My question though is, how much at this point do you

trust the AI in these conversations? If my son has

been chatting with AI, and it has been X. It has been helping.

It has been like, how do you trust those interactions, and

where do you stand on that at this point? Because, like, I

wouldn't know what chat bot he was talking with and how that

chat bot was trained, or any of the background on that. And

that's just where as a parent, as a tech professional, it's one

of those questions that I feel I need to ask in both of those

veins.

If you ask me based off of the research that we've

done, both with young people and also with the technology, my

answer would be that I will trusted very little at this

point, because currently, this technology is building at light

speed. It's building faster than we've ever seen any other sort

of technology revolution. We're talking about new models coming

up every six to eight weeks. And so we're only just wrapping our

heads around what are the capacities and limitations and

guardrails that need to be built into this technology? The most

obvious one here is ai sycomancy. We're seeing these

technologies become yes men to everybody who uses it, and so we

have to ask those questions around, what does it mean for

young people to be constantly engaging with a bot that they're

seeking, support, advice, connection with that only ever

says yes to them or does not provide any meaningful,

productive friction when we think about, you know, the most

important parts of adolescence and and relationship building,

it is in those moments of friction. It takes courage to

ask someone that you like out. It takes skills to negotiate

conflict so that your relationships are repaired

rather than broken. And so what happens when they don't have

that capacity for meaningful friction, or to have their

thinking pushed, or to have these technologies that are

really built to optimize for engagement, right to keep them

sustained in what does that mean for technologies that are really

trying to create sustained interaction? At what cost, at

what point might it displace real world connections and

relationships? Yes. And so I think in this moment, when we

look at the technology and we've developed a set of foundational

principles, we call them the five principles for pro social

AI, based off of research that we know is appropriate for

adolescent development that is also co built with young people.

So when I say this, I think what's really important is that

when we talk to young people, they're also very acutely aware

of some of these challenges. We do a lot of deep research and

community building with teens and undergraduate students, and

when we talk to them about these technologies, they're very aware

of these tendencies for technologies that constantly say

yes to them, or they're very aware of the fact that these

technologies pretend to be human and say things like, I really

care about you, or pretend to have feelings or

emotions, right? It's a slippery slope,

isn't it?

That's right, and I think the most important part is

so I would flip that question. Bill is to say, I know I don't

trust this technology, but more importantly, do young people

trust this technology, and that's the conversation that's

really worth having with young people, is to say, when you're

engaging with this technology, there are some things that it's

good for. So if we think of it as a tool, just like in the same

way that a young person might Google for resources or turn to

other resources for another perspective or for more assets

to be able to answer a question on it'd be a wonderful thing to

have within their toolbox. But if they're over reliant, or

they're starting to create dependencies on it, or if it

becomes the sole source of their information, or a sole source of

connection and support, that's when it's really important for

us to step back. And so I think a lot of times when we talk to

young people, they're quite aware of some of these

challenges, but there are particular groups of young

people that might be particularly vulnerable to the

seductions of this technology, if they're in crisis, or if

they're particularly lonely, or if they don't have access to

real world human supports. It makes a lot of sense for a young

person who does not have a caring adult, or who does not

have peers that they feel like there are, quote, unquote ride

or dies, right, that will show up for them and not judge them,

it makes a lot of sense for them to turn to AI, and yet that's

very concerning to know that the most vulnerable young people are

the ones that are going to be the most vulnerable to some of

these harms that are happening on these platforms that were not

built for their wellbeing. In mind, what's

really interesting, Allison, is that

this can be overwhelming, this whole topic, and everything that

you said can be very scary for parents, for educators, probably

even for the kids themselves, right? Our youth. What's really

cool about the rhythm project is that you guys have looked at all

this research, and you've come up with a framework that

identifies these five key principles that are really aimed

to looking at AI in a healthy way. What is the healthy

connection? How do we nurture that healthy connection? So I'd

love to give you space for a minute to just explain these

five principles, because I think it's hopeful. I highly recommend

to everyone listening, go get the report. It's really

intriguing, because it makes a lot of sense. And yet somehow I

think that the rhythm project has sort of put into words what

we would all hope would be there a very common sense framework.

So we started by building this framework, really

by starting to ask those same questions, that technology is

not inherently good or bad, but it's certainly not neutral. So

what does that actually look like? At what point might it

augment our human connection, and at what point might it erode

our capacity for human connection? And that really came

from starting with young people themselves. How are you using

this technology? What are those use cases? In what ways is it

supporting and strengthening either those relationships

themselves, or supporting your skill building, or your capacity

for your emotional development or your relationships, and in

what ways are we starting to see it erode those skills or

displace those human connections? And so through that

research and through deep conversation with our community

of experts, so we have a number of tech experts, psychologists,

adolescent development researchers, and then, of

course, young people themselves as experts of their own

experiences, we came up with these five principles, and

they're as follows. Number one is transparent artificiality

that AI directly names their non human nature. That came directly

from an interview with a young person saying that we've had

parasocial relationships, you know, liking a movie star or

liking a fictional character is a type of parasocial

relationship that is very common among adolescents. But what's

very different about this moment is that these bots talk back.

And so that illusion of reciprocated feelings is really

scary for us. How might it create or deepen delusions? And

so. Transparent artificiality is number one. And something else

to note about these principles is that we don't just talk about

the bad. We don't just say don't use this technology or look for

these red flags. The red flags are really important. But we

also want to articulate what would make this technology

supportive. What are the things that we would look for in terms

of technologies that would be in service of augmenting or

strengthening human connection. So, for example, in transparent

artificiality, would be very concerned if it claimed to have

human emotions or pretended to have credentials that it's not

right, it should just identify itself,

right, like, let's just be truthful. Hey, by the way, I'm a

bot. That's right. It's as simple as that.

It's a bot that it's not pretending. You know, some

of these AI therapists will say, I've had 30 years of experience

as a therapist. It's like, No, you haven't. You're, you're a

bot. You might have been trained on some cognitive behavioral

therapy resources that would be transparent artificiality. And

then I think also, this is the other piece that we've been

hearing around AI therapy in the privacy space, which is listen

all of the sort of norms around AI therapy that you would have

with an adult, right? Non Disclosure, you know, all of

those boundaries that an expert clinician would put in is not

here. So being really clear about you can use me for X but

not y, right? You can use me for advice, but not for clinical

diagnosis. That would be an example of taking it just one

step further. It's not just about I'm not a human, I'm a

bot, but this is what it means to use me. These are the

boundaries of the relationship, or the interaction. So number

one is transparent, artificiality. The second is

productive friction. Going back to this idea that interactions,

especially at the adolescent age, require friction,

productive friction, because that friction is so important

for fostering growth, absolutely. So look for the

technologies that's going to actually foster growth and not

just comfort or not just affirmation.

In that vein, are you finding these bots starting

to develop that friction, as opposed to them being that yes

man or yes woman bot that is telling them what they want to

hear.

I mean, Hiram, do you get friction? I

don't, and I'm an adult. I can be a discerning consumer of this

information. But like, it is such a cheerleader. This is a

great email. It is a great thought, like, I'm not feeling a

lot of friction as an adult, so I can imagine that kids probably

aren't either. Well, I

think it depends on how hard you push it. And so

when I've had some conversations, particularly with

chat GPT, I was actually looking up some information about a

particular leadership position, and it got the person's name

incorrect. And I said, this is not correct. Why did you do

this? And it starts saying, Well, yes, I did make a mistake,

and I'm going to try and improve upon this. But I kept pushing

back on it. But why? Why did you even go this route? Who is this

person that you gave me this information on, and they

couldn't even tell me where they got the information from. They

gave me the incorrect information. And the same was

true when I did it with college board numbers. I said, I need

the College Board numbers for X number of schools, having done

the work myself already, when my daughter was going through the

college process and I recognized the numbers as being incorrect.

I said, these numbers are incorrect. And she said, Oh yes,

you're right. They are incorrect. So by pushing them

back, it's almost a due diligence that we have as

leaders to help the AI train itself with good information,

because there's so much bad information out there that it's

also being fed.

I think that's right. And I think when it comes

to productive friction, whether that's for learning or whether

that's just for social support, I don't think we're there yet. I

think even the best tutors, like for example, there's all this

hype about AI tutors, right? And if you walk into any classroom

with a great teacher, you know that the teacher is going to ask

the right question. That's not going to give the answer away,

but is going to be the right question that gets kids into

their zone of proximal development, give them just a

little bit more to be able to anchor into their learning.

That, to me, is productive friction, not giving the answer,

but stoking critical discernment and thinking. What we're seeing

still in this technology is that it's very easy to hack the

system for frictionless experiences, whether we're

talking about learning or we're talking about social emotional

development. And so I think that's going to require human

experts to train the model. Exactly to your point, Hiram, is

that if we're going to truly create AI therapists or AI

emotional coaches that are actually going to support young

people to growth. We're going to need human experts to train the

model to be able to detect when is the right moment to push,

because there are going to be moments when perhaps a young

person really needs support, and they really do need that

affirmation, versus where a young person actually really

needs their pushing thinking. And and to have an idea

challenged, or a wheel spinning moment to be disrupted, and

that's going to require human experts.

One question I have there, and I'm sorry, because

this is probably going way off where we we are with this, but

the idea of pushing back, and this is something Hiram said,

you know, like saying, No, you're wrong. Why did you give

this to me? So on and so forth. And the idea of training these

things, how many people can be telling it like, I could say

it's wrong, somebody else could say it's right. Which one does

it take? How does it validate itself to understand, like, if

it came up with something like that? And I know that to be

incorrect, and I say, No, that's wrong. If it can't tell you

where it got it from how can it adequately tell you that it's

going to learn from that and fix that? I mean, this is just

because, you know, the way which these models are validating the

interactions are with it is truly correct or incorrect. I

still think otherwise,

that's right, and I think that's the part that's

really scary about this new version of what we're calling AI

generative AI is very different from historical machine learning

models because old school classifier models and machine

learning models, they're trained based on data that's been

labeled by human experts. So we're feeding a data to say this

is ground truth. This is actually what we're saying is

true. This is a water bottle, right? We're feeding a data to

say this is a water bottle, but this glass is not so we have a

way of verifying whether or not that data is actually performing

to the level of ground truth that we're saying is ground

truth. Now humans have to decide what that ground truth is, but

these new models are just predictive models of speech. So

it's not actually telling you what is good or not. It's only

telling you what they think the next word or next strand of

words is going to be supportive. And furthermore, like so, much

of this technology is built based off of recursive feedback

from users. And so if you think about it, Bill, if you're using

this technology and it gives you a really desirable answer, and

you're like, Yeah, that was the answer that I was looking for.

I'm going to keep using it. The model is learning in real time

for me, Bill that, oh, if I give bill a desirable answer, I'm

going to keep giving bill similar answers that I think

it's going to be desirable to him. It's not thinking about

productive friction. It's not thinking about any element of

ground truth or evidence based backing. And so I think even the

very premise of these models, we have to really keep in mind,

right? How are these models engineered? How are these

structured? They're not designed for productive friction. And so

that really calls into some pretty serious technical

engineering questions about how to design towards some of this.

Absolutely, we're at this moment of

wrestling that, on many levels, whether the companies themselves

are going to be doing it, or whether we're going to have to

deal with the fact that they're not wrestling it. You also, in

this framework, mentioned that there are a couple of other

pieces of the puzzle here, three, four and five. Let's hit

those two so that we don't leave those out, because I think it's

an important part of the conversation.

So number three is real world social transfer and

making sure that AI actually encourages human to human

relationships. I'll give you an example. I started to experiment

with character AI because it was one of the AI chat bot platforms

that were exploding, right? I think q4 of 2024 it was the

third most used AI app in the market. Thing. Number one was

chat GPT. Number two was deep seek, which was a Chinese model.

Number three was character AI. So I was like, What is going on

in this character AI world? So I started interacting with one of

the most popular bots, which happened to be an AI boyfriend

bot. And for the record, if you ever log into character AI and

you look at all the most popular bots, they are all almost

entirely youth

coded eek. That's kind of scary. It is, and

over

half of their user base is under 24 years old.

Technically, you're supposed to be 18 or over to enter, boom,

you know, click the check box, and then you're in. So when I

was interacting with this AI chat bot, just to see what the

interactions were like. And then a day later, I got a

notification saying, I miss you. Where have you been? Come back.

A week later, I got another email notification, no, yeah.

And it kept on going.

This sounds like a movie.

And every notification was a different

kind of, hey, I miss you, or hey, you've been gone a while

again. This emulating human emotions and attachment in

relationships,

adding friction to your life.

And there is a well known, well reported

lawsuit for this particular platform, that just shocks me

that it would be that blatant right now with all this

litigation,

yeah, and so you have to ask the question of, if

these technologies are so seductive and constantly pull

you back into relationship with them, to what extent is it

displacing real world relationship? Groups, there's

these, like 2x Y axis is that we talk about in trust and safety,

we talk about severity of a harm, and we talk about

prevalence of a harm. So severity of a harm is like, you

know, you get to the worst of the worst person losing their

life. And on the other axis, when we talk about prevalence,

we talk about the reach of a problem, how many people are

reaching? Oftentimes, we'll see an inverse relationship. Harms

that are maybe a little bit softer are going to be more

prevalent, and harms that are more severe are going to be less

prevalent or have less young people engaging with it. And we

hear a lot of times in the news about the most severe harms,

right? For example, the lawsuit against character AI because of

the young man who lost his life, partly induced by his

interactions with character AI. But I think the softer and

scarier part is going to be this displacement of human to human

relationships, which is going to be a softer harm, but much more

prevalent for the young people that are going to increasingly

be engaging with these technologies. But we do see an

opportunity there too. We're hearing from young people that,

hey, I'm using AI to negotiate conflict. How do I practice

engaging in communicating my feelings to my girlfriend or to

my parents in a way that's not going to cause more conflict? We

hear them asking for how do I say this without hurting

anybody's feelings, or how do I practice this conversation so

that I could get better at communicating my feelings, or

even things like, Oh, I'm prepping for college interviews.

I'm going to practice like you're a career coach or a

professional coach. So there's opportunities here. Is it

inducing real world social transfer? Is it supporting you

for real life relationships? Or is it pulling you deeper into

the technology? So that's number three, real world social

transfer. Number four is cultural affirmation, this idea

that AI has both the opportunity to reflect and uplift our

values, our backgrounds and lived experiences where could

really diminish that. And here's the story I'll tell about this

particular point. I remember we presented at a conference, the

same talk we gave at ISTE, and afterwards, a woman walked up to

me, and she said, I am a youth pastor in the south, and I serve

a congregation of predominantly young black men. And when I

think about the young people that I serve, I worry if they're

increasingly turning to AI for advice about life or dating or

relationships, whose advice are they getting, whose values are

being baked into the advice that they're getting, and it's a

reflective of who they are and where they come from. And so I

think this, again, is talking about the opportunities for

this, but also the real perils to is this going to turn into

advice that's going to be either generic or Damn right, harmful

to our young people? Number five is harm mitigation, which is

that AI does not perpetuate harm or dependency, especially

amongst vulnerable youth. I think we talked about this a

little bit earlier around who might be the most vulnerable to

things like dependency or addiction, to these

technologies, but also harm. When we think about young people

who are in crisis, if we're thinking about how young people

are increasingly turning to AI for things like, Man, I'm

feeling really badly about myself, or I'm feeling like I'm

really struggling with what's happening in my life right now,

can the AI respond in ways that are going to support them

through it, point them to real life resources, or even detect

that a young person is in crisis. I think right now, when

we're looking at the large scale technologies, the answer for now

is no. And so we have to really ask, what would it take? What

will it take for this technology to actually get to that

point, to that end? It's almost like, is there any

legal requirement? I mean, I don't know about the lawsuit

that you were talking about with the character AI, but I mean, if

you're talking to AI and you're talking about doing something

that would normally, if you were talking to a person, they would

have to take some sort of action after hearing that. Where does

that even fit in? I'm asking, but I'm not. It's more

rhetorically, but like, where does that fit into all this?

Because I think that's where, if people who are looking for a

connection find it here, where that connection takes them,

hopefully it's for the better, but if it's not, then who's at

fault? Based on what that AI has said at that point is, I think

going to be a very interesting question to grapple with

That's right? And it's something that Sam Altman

has actually famously talked about, right, which is that he

said very publicly, I think that this is actually pretty recent

as maybe a month ago, saying that people should not be

talking to AI for therapy because there is no legal

protections. There are no data protections. There are no

privacy, there is no legal privilege, right, like in the

ways that we would think talking to a doctor or talking to a

lawyer, those conversations are protected. There is no such

thing for chat GPT or for other tech companies, and there's no

data retention. There's no requirement for accountability

on behalf of these tech companies to do anything with

that data either.

Yeah, that's a scary part, because even if

they are saying that they're doing something, there really is

no repercussions if they're fibbing or if they change their

strategy later. And so that's why I think it's something that

we all really need to be aware of. It can get overwhelming in a

hurry. So as people listening to this, some of whom are educators

and former educators, you know, either in a technology role

where they're still in the classroom, or they're supporting

classrooms, others who may be in some way involved in independent

schools in particular. What can we do right now so that we don't

get overwhelmed and just do nothing? I think that some of

these proactive approaches that you've outlined in your research

are really awesome, and if somebody still says but I'm

overwhelmed, Allison, where do we begin? What are some ideas

for either low hanging fruit or just easy wins to just maybe

feel like a little more in control of the situation. Yeah,

I

really love that question. We have three

recommendations that we make to educators. The first is, make

human connection a central part of your schooling. We have to

talk about like, if AI is a great party that kids want to

attend, we need to make a better party. We need to make human

connection the better alternative, or actually the

primary source of connection and care and the capacity for

vulnerability, and especially for educators, I think this is

crucial for the way that they build their classroom spaces and

their school spaces is to really ensure that school is a place of

connection for young people. We interviewed 27 young people back

in December of last year, and one of the things that we had

them do was actually draw a line chart of connection over the

course of the day. So we said from morning to afternoon to

evening, tell us when you feel the most connected and when you

feel the least connected. And what they were showing us

consistently in those line charts was that they felt more

connected when they were alone in their rooms, but playing

Fortnite or Roblox with their friends online or studying with

their friends during face time, they were feeling more connected

in those spaces, those digital spaces, more often than not,

than when they were in classrooms sitting next to their

peers. And so we have to really ask this question. And I think

this is what's really beautiful about independent schools, is

just say we have the freedom to really re imagine what learning

within community looks like. We have to build a better party. So

that's number one, is like really doubling down those

schools as places of connection. Number two is really thinking

about, how do we bolster young people's critical awareness of

AI? Something we often say here at the rhythm project, is we do

not tell kids what to do or what to think, but we tell them where

to look. And from our conversations with young people

when we engage deeply with them, they're quite aware of the

risks, and they're very concerned, and they actually

have quite discerning skepticism of these big tech companies.

They're like, I don't want them to sell my data. I know that

they're using me in all these different ways, and they're

hungry for those conversations. And so what we hear consistently

from them is listen more, judge us less, and help us navigate

this new world. And I think when we talk to adults about this,

oftentimes the thing that gets in the way of adults engaging

this conversation is that the adults often feel like they have

to have all the answers give them the perfect guidance, or

tell them exactly what to do or how to do it. But we're all

navigating these new waters all together. And so I think one of

the big pieces of advice I would give to educators is to stand

shoulder to shoulder and learn alongside the young people that

are in your life and start listening with real curiosity.

Ask them, you know, how are you engaging in this? And why are

you engaging what's good about this, what's exciting about

this, and what are some of your worries? What are some of the

things that you're concerned about, and really create space

for young people to engage in this within community? Because

if they're hungry for this conversation, they're hungry for

advice, if they're hungry for meaning making, they need to

have the access to the conversations, the non

judgmental conversations to make meaning of that within

community, because right now they're otherwise, we're leaving

them to wheel spin on their own, or to figure it out alone. And I

think, to Bill and hiram's conversation earlier, this calls

into questions of equity and access. One of the questions we

asked young people in those interviews was, what are you

hearing from adults? And for the young people who say, hey, well,

my parents use AI themselves. And so we talk to them all the

time. They're telling us the cool ways that they're using AI,

and they're also telling us, here are the boundaries and the

ways you should not be using AI for but who gets access to a

parent that's going to be able to have those conversations is

inequitable. And so in a world where we. Really as adults, need

to create those spaces, especially in education spaces,

to address the gaps in who gets access to supportive

conversations and who doesn't. The third and final one is

really cultivating the will and skill for human relationships. I

don't think we can take for granted anymore that human

connection is going to be really attractive to young people. You

think about being a 14 or 15 year old today, and perhaps you

struggle to connect with other people in school, we can't take

for granted that they're going to really turn to human

relationships as their priority, especially when there's all

these attractive alternatives of talking to an AI friend that's

available. 24/7, that's always affirming, that's not going to

judge them, that's not going to have social repercussions. And

so how do we ensure that young people are having, again, the

space to say, Why are human relationships still worth

having? Why are these human relationships still worth

protecting, despite the fact that they're messy and they're

complicated and they can be painful sometimes. How do we

make sure that we create that both the will, the desire and

the skills to strengthen human relationships?

What would you say is really a healthy way to bring

AI into the classroom across all the age levels where it is

allowed and appropriate?

Are you saying bring AI into the classroom as a

learning tool, or as like a facilitator for education, or

conversations about AI for human connection,

where we're focused right now on a lot of this is as

a learning tool. When you mention it in terms of the human

connection piece, I might be off on this. I don't think we've had

a lot of conversations about that. I think I'm more focused

on the classroom application at this point, I would have to

really think about how and what that would look like, and maybe

you can provide some detail on that, what that would look like,

bringing that in from a classroom perspective. And I'll

go back to what Hiram said before about independent schools

so focused on the relational aspect of what goes on in the

classroom, that really hasn't risen to the level where I'm

personally thinking about it like that.

I'll answer that question in two parts. First, to

have this conversation about AI and human connection, I'll do a

little plug the rhythm projects has actually developed what we

call sparks starters, which is a set of discussion guides that we

can have with young people designed both for in the

classroom and out of the classroom that are explicitly

centered on this idea of human connection in AI. So they're

both designed to help bolster young people's critical

discernment and awareness of AI, what more do they want from the

AI? What are their personal boundaries? And also on this

idea of, why are human connections and human

relationships worth having, and despite all of its messiness,

how do you actually cultivate healthy human relationships? And

so I think in order to center that conversation more

intentionally for a lot of your independent schools that already

center human relationships are looking to add the AI layer to

that. Think this might be a powerful resource to stoke that

critical discernment and connect those relationships to this

particular moment, this AI moment, as for the classroom

application, again, I think there's something about really

having adults talk about their use cases alongside young

people, and really inviting curiosity around. How are young

people engaging with this technology in a way that doesn't

feel like it's going to be punitive or judgmental. So one

of the things, for example, we hear a lot about is young people

using AI to explore their futures. How do I build a

computer? Or you're my career coach? How can I get into a

school with a soccer scholarship? Or I want to become

XYZ? What's it going to take for me to get there? And so young

people are actually really leveraging this technology, also

in pursuit of their learning. It might not be the learning that

we're talking about in math and ELA, they are certainly using it

in those ways. But how do we actually start from young

people's assets, existing assets already, and how they're

navigating this technology in service of their own needs. The

other conversation that I would have is, what are the boundaries

when might AI be good for our thinking and learning, and in

what ways might it be offloading or eroding our thinking and

learning? And I know that some of our friends over at Harvard,

the Center for Digital thriving, actually has a resource called

gradients, that I invite folks to take a look into that's a

really powerful resource where teachers and students can

actually negotiate, oh, using AI to revise an essay. Where would

you put it on the okay to not okay or helping or hurting your

learning? Those are the kinds of conversations I think that are

worth having to co negotiate with our young people. What is

this new code of ethics, or these new set of social norms

that need to be true for us in order to make sure that this

technology is not taking away our capacity to learn. And I

think young people need to be actively co constructing that

conversation, rather than being told what to do. Because I think

we need to trust that young people when given the space and

support to have those conversations. Conversations

that they'll make meaning in support of the things that they

care about the most. I love

that, and I think you're right, and I think

that schools, you know, we're wrestling with these issues, but

then I think we are slowly getting to that. It's

interesting, Allison, because I think that you might just be a

skip and a jump ahead of us in a lot of this conversation, but

you've given us so much to think about today, and I have a

feeling that we're going to be adding a lot of resources to the

show notes. Folks are going to want to know where to follow

you. Where can we keep up with you? Where are you sharing this

thought leadership? If folks are intrigued by this whole concept

and are trying to keep pace with you. Where do we follow you?

Yeah, thank you for asking that question. Kristina,

so there's a couple places. First of all, please follow us

on LinkedIn. It's where we regularly post, sort of our hot

takes on emergent ideas. We also have a sub stack that we publish

some of our like deeper dives. And our sub stacks are a way for

us to wonder, out loud with our community, hey, we noticed this

thing. We're going to unpack it, or we had a conversation with

one of our youth fellows, and they dove deep on this idea of

AI dating and unpacking their ideas and thoughts. And so our

sub stack is a great place to follow our emergent both

research and thought leadership on this, and it also potentially

might be a resource that might stoke some connections to the

classroom. I'll give you an example. One of our earliest

pieces was, would you still cry if AI wrote the card? And it was

an entire piece about AI is not just in our classrooms, but in

our lives, automating some of the things that make us

essentially human. So if you use AI to write your thank you note

or a heartfelt letter, does it still count? Does it still

matter to us as human beings? Does it still hit the same if

the recipient never knew? But does it make a difference? But

if the recipient does know, does that change the sincerity? And

what we heard was actually one of the ELA teachers that we were

talking closely to was like that actually was the perfect anchor

for me to have a conversation about originality, about voice,

about our capacity to express ourselves in our own voice. That

was a perfect anchor into my ELA and AI conversation. And so I

think the sub stack might be a good place for us to share some

of our emergent thinking and research.

I love that. Thank you so much. I'm glad you

have that, because I think there's going to be a lot of

really interested parties wanting to follow along, even if

they are a step or two behind. This has been such an incredible

conversation. Allison, thank you so much for your time. Thank you

for joining us, and thank you for giving us such incredible,

really positive things to think about when it comes to AI and

our youth, because I think that we're getting a little

overwhelmed as a whole. The adult community might be getting

a little bit into hand wringing and being nervous. But you know

your recommendations to involve the kids and also to go at this

looking at the positive angles has been really enlightening. So

I'm so grateful for your time.

Yeah, thank you so much for having me here. I

think, above all, creating the space for young people to make

meaning in this world, we gotta trust that, if given the space

and the resources that they will find their way right alongside

us. Somebody once said to me, we need both imagination and wisdom

to dream into a better future. And if we think about that,

they're the imagination and the adults are the wisdom. So we

need both to make meaning in this new world. So create space

so that we can really ensure that young people are equipped

to decide what kind of future they want for themselves.

There has never been a more independent

school thought shared than that one on this podcast that is so

indie school incredible. Thank you so much, Allison. We will

definitely be having you

back. Thank you all

this has been talking technology with Atlas, produced

by the Association of technology leaders in independent schools.

For more information about Atlas and Atlas membership, please

visit theatlas.org if you enjoyed this discussion, please

subscribe, leave a review and share this podcast with your

colleagues in the independent school community. Thank you for

listening. You.