Evan Harris of Pathos Consulting Group addresses the escalating risks of AI-generated deepfakes and companions within school communities. The discussion highlights the psychological tactics scammers use to bypass critical thinking, the limitations of technical detection tools, and the urgent need for schools to update policies and student safety protocols.

• Pathos Consulting Group - The organization founded by Evan Harris focused on deepfake safety and crisis readiness for schools.
• Reality Defender - A digital forensics tool mentioned for its use in identifying AI-generated content.
• KnowBe4 - A cybersecurity awareness platform discussed in the context of phishing and deepfake training.
• "Fake" by Perry Carpenter - A recommended resource for understanding the science of deception and human nature.
• Presidium - An organization cited for its data on the rise of youth-to-youth abuse within independent schools.
• Adaptive Security - A company providing cybersecurity training for high-stakes organizations.

Paul, 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, Christina Llewellyn.

Hello, everyone, and welcome back to

Talking Technology with Atlas. I'm Christina Llewellyn, the

President and CEO of the Association of Technology

Leaders in Independent Schools,

and I'm 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.

Gentlemen, we are live from the stage in

Columbus, Ohio, for the 2026 Atlas Annual Conference. How has

your day been so far? This is the family reunion that we look

forward to all year long.

It is indeed. I am doing great. I am doing much

better than I've been doing the past two days. Been a busy

weekend working here with the board, getting everything set

and ready for the next phase of Atlas. So, it's been a fun

weekend to start, and looking forward for the next few days.

Absolutely, it's two very busy days with the

board, but we are in a great place. We are so looking forward

to the next two days, we just finished a wonderful keynote, so

we're excited to get in that conversation.

I do have to say, for those of you who are

listening after the fact and were not with us live here in

Columbus, we just had a really emotional, incredible general

session where we did a lot of welcoming and handing over of

the reins as I transition out of the CEO role and into my new

organization, but I have to point out for our listeners that

I did finally identify my zombie melee weapon of choice,

outstanding,

and I would be remiss in saying that Hiram and

I, we were doing a little deep research in terms of what that

actually looks like, what it represents. Just,

yeah, I mean,

it was the right choice.

Look, I asked AI, knowing what you know

about me, which is probably scary, what would my weapon be

for fighting a zombie, and it gave me a fire ax.

Nothing better.

Okay,

it made perfect sense.

Okay,

so watching your ax throwing prowess.

Oh, I definitely can kill the ax

throwing situation, so

indeed.

Yeah, well, there we go. So, it's finally on

record. I figured I better figure out what my weapon was

before we finally wrapped up this era of me being on the

podcast, because everybody needs to know what my weapon is. So,

it is apparently the fire ax. So, we'll have to link that

whole thing in the show notes. Awesome. Well, we're really

excited to be here with Evan Harris. Evan, you just got off

the stage, and we're no rest for the weary. You are right here

with us, recording the podcast. Thank you for an incredible

general session. Please introduce yourself to our

listeners.

Yeah, thank you so much for having me. So, my Evan

Harris, I'm the president of Pathos Consulting Group, and we

help schools with deep fake safety, and so that looks like

helping schools with policy with crisis readiness plans that

comes off in the form of a tabletop simulating a deep fake

incident at the school, as well as preventative education for

every major stakeholder group, from leadership teams to board

members, parents, faculty, staff, all the way down to

students, both at the upper school and middle school level,

and so this work has now been going on for a few years, and

we're learning a lot about what's working in terms of best

practices on all these different areas. Deepfakes is a lot to

take on all at once. It's a fast-moving target, but we sort

of feel like we're starting to get our arms around it again,

those kind of major tent poles, the policy, the crisis

readiness, and the prevention.

So, what's really interesting is that you

all could have heard a pin drop in this room when you were

speaking, because the deep fakes have gotten so, so good. How do

you even describe a deep fake? What is

it, without over complicating it? It's an

AI-generated imitation of reality. It comes in the form of

an image, a video, or a vocal clone, where someone essentially

creates a digital twin of your voice, and I'm thinking too

about imitations of human emotions, also in the form of

like AI companions, and some of the ways in which we're seeing

that impact our society, and especially our kids. But the

kind of main thing that I try to emphasize to people is that even

if you're very smart, very discerning, very media literate,

you can be fooled by these things. I have been fooled by

these things in the last six months, the line between what's

real and what's fake is sort of disappearing before our eyes.

This impacts a number of different areas across school

life, so we have to start getting ready before something

happens, not afterwards.

It was so interesting because you put some

great examples up on the screen, and basically asked the

audience, left or right, is this the fake one, or is the one on

the right the fake one, and we failed the test. It was maybe a

5050 or 6040 but you said that kids do a better job than adults

in detecting the fake one.

Yeah, so this was fascinating, because I thought,

oh, I'm in front of a room full of tech leaders, this will be

the audience,

we let you down.

I was like, this isn't gonna work here. There,

and then, no, it was basically 5050 split. I've done this same

test, the spot, the deep fake test, with audiences of kids now

across the country, even kids as young as fourth graders. It's

like clockwork at school after school, 90% of them are getting

this right, and I emphasize to people, they're not getting it

right because they're smarter or more media literate, it's

because they're swimming in a sea of AI swap, and they're

developing a sixth sense for this. This is actually not great

news. It's breeding overconfidence. Kids think that

they don't have to learn about deep fakes because they already

can spot them, but that has an expiration date. The technology

is getting more and more realistic, and I share this also

to impress upon schools the work of telling kids about deep fakes

has been done for you. You now have to inform them and protect

them. Do your deterrence messaging about deep fake abuse,

and also let victims know that if they come forward, they're

going to be supported. Not talking about deepfakes is not a

solution. They're already ahead of us on this. We're the ones

that are behind. So, I mentioned AI companionship. I always root

this, or try to root this, in a sense of empathy for kids, which

is that if you're a kid who is isolated or vulnerable, you're

going through something really difficult, a moment of crisis.

Having a companion, an AI companion that is always

available, never judgmental, and agrees with a lot of your ideas

is going to seem very attractive, and I think we can

recognize that and show empathy for that. Here's the problem:

there are a couple of issues with this that kids have to be

aware of. The first one, these are not mental health

professionals, and they did a study of AI companions and found

that they only respond appropriately to a child in

crisis 22% of the time. If you think about what a real

therapist does, they push back on your bad ideas. AIs are

designed to be sycophantic and to have you engage, and so

they're not designed for that task. And real kids have

self-harmed because an AI told them it was a good idea. Another

thing that I worry about is a child who's learning patterns of

what it means to be in a relationship from one of these

AI companions and have that transfer to their real-life

relationships. Middle school, high school, it's a messy,

awkward time to be learning these skills, but I'm worried

about not just what kids are opting into with these

companions, but what they are opting out of by not learning

those hard skills that are again messy, awkward, but necessary.

They form the foundation of what we use our whole lives to relate

to people. The last one are the privacy risks, if you were using

AI as a confidant, a therapist, a romantic partner, you were

probably inputting very sensitive information without

fully understanding where is it stored, can it be resold, does

it ever get deleted? Kids don't necessarily think through those

things, and so I worry about AI companionship from a couple of

different angles.

So, before I let you guys get into your

questions, and I know Bill and Hiram, you have some, I want

your reaction. You were in the audience, first of all, quick

question, did y'all get it right? Did you pick the right

fake? No, yeah, I didn't either, Bill. I'm gonna admit I picked

the wrong one. So, Hiram, you got it right.

Yeah, I could tell the perfume one, it's really

difficult to hold three bottles,

yeah,

yeah.

So, what was your reaction overall to Evans

chat? Were you prepared for it? Are you more worried? Are you

scared? Are you concerned?

I mean, I want to target one of your statements,

which I think will perhaps push this conversation that the goal

of AI fakes really is to short circuit critical thinking with

emotion that really struck a chord with me.

Yeah,

then all of the examples you provided afterwards

were all stunning examples of that very statement. Can you

tease that out a little bit more for our audience?

So, I think you know, when it comes to deep

fakes, so much of this feels new, and so anything we can do

to look back at the science of how misinformation spreads, the

science of kind of how the human brain works puts us on much

firmer territory, because while the technology changes, the

human brain stays basically the same, and so as the technology

gets better, I'm teaching more and more about that psychology

piece, and there are a couple things that scammers do that are

psychological in nature, fake urgency, fake emotion, appeal to

authority, confirmation bias, wish fulfillment. These are all

very, very studied, but the thing you're alluding to is this

idea that when we encounter emotionally charged material, we

are more likely to believe it and more likely to share it, and

it doesn't matter what kind of emotion it is. It could be

something that's designed to enrage and inflame, be divisive,

but it could also be a baby or a puppy or a wedding, and you'll

notice in sort of the deep fake content that's getting

generated, they seem very aware of this, they're spinning up

very emotional material because they know again that it short

circuits critical thinking, and we're more likely to believe and

share.

Yeah, listening to it, there were a lot of things

that really struck a chord, and it was interesting for me. After

you concluded, I went back because I came here with a team

of five from my school. I checked in with them just to see

where they were, and there was a lot of questions, like where are

we on this, what's going on with this, and even when I was

sitting next to Hiram listening to you speak, he was alluding to

some of the things that he was doing at. In his school, as it

relates to some of the voice fate, I was sitting next to Eric

Hoffman, and I jokingly said to him, "This is why I don't answer

my phone anymore. Why let Siri answer it, screen it, because

I'm not going to pick it up unless I know that it's from a

trusted caller. At that point, the conversation that I had with

my colleague at the time is like this is just a growing piece of

what we've been dealing with, with regard to, like, I call the

head of school email, that you get, and it's even gotten to the

point of texting, you know, when I get them, when I see those

come through, I always send out, like, a message explaining and

pointing to the reasons why these things are off, and it

just speaks to me to the amount of work that we need to continue

to do in our due diligence around the cyber risks and the

risks that we take on to protect people and how we can think

about that in your example of the father that thought their

kid needed to be bailed out exactly it's those types of

things that I think about it how do we talk to our faculty or

employees about it from an institutional perspective, but

also from a personal approach, because I think if you can land

it with them in that personal realm, it carries through in the

work that we're all trying to do at school, so it's a two birds

one stone type thing. If we can hit this right,

I've noticed in doing sort of faculty

facilitating their sessions, a lot of people take off their

teacher hat and they're a parent or a child, right, and so

they're thinking about protecting their kids, and

they're very often thinking about protecting their parents,

and so when I talk about needing a code word, yeah, that's for

your business office, but it's also for your family, and I

would definitely include your parents in that, but that idea

also of second channel verification is another sort of

important step, and I think teaching the psychology of

because these scams, they repeat the same sorts of tricks over

and over again, sort of anecdotally, I don't know how

well documented this is, but I've done enough of these

sessions to have heard a lot of people who've been scammed this

way, or they attempted to scam them this way, one of the sort

of odd things that they'll do with this particular scam you

alluded to is that it's not just that someone you know has been

hurt, it's often that they've broken their nose, and they do

the broken nose thing because they want you to attribute any

discrepancy in the voice to the broken nose, and so they've

really thought through all these little things, but if you're

kind of aware of the technology, you've been exposed to it, there

is a bit of an inoculating effect to that.

It's really interesting. So, not long ago, I

instituted the code word thing with Atlas' staff, and it just

reminded me of being like a kid in the 80s, right? Like, we were

always told that if your parents couldn't come pick you up at

school and some stranger danger to avoid the stranger danger,

they would send somebody in, and there would be a code word,

right? And so I don't know why that came to me a while ago, but

we have one Atlas that we've never put in writing. We've only

expressed that code word on a staff call because it's been

necessary. Like, Christina, did you get a new cell phone number?

No. And so, like, the code word is already baked into how we do

business continuity and safety protection at Atlas, but I never

thought about doing one with my family.

Yeah, and what you're describing, by the way,

is making it part of the culture, normalizing it, making

it normal to ask, so it's not awkward, but just like it's just

a normal thing to ask for. I remember I'd worked with a

school and I emailed them afterwards because something in

the invoice needed to change, and then they called me

immediately afterwards to verify that it was really me, and I

thought, oh, the training worked, they listened, they

immediately applying it perfect.

So to that point, one of the questions you

mentioned early on that I'm really curious, because often in

the work that we do as tech leaders at our schools, we're

asked to come in and either to verify or to help do like a deep

dive on that forensic side of things, which you mentioned

early on in your talk, and I was just wondering, like, if I'm

called in to do some of that work, I know what tools I have,

if it's like related to Google, if it's related to another

service that we have some level of admin access to it, but how

do you talk to schools about, and you did list off a few of

those things, but are there any ways in which schools can

leverage forensic tools to understand that, because so

often we're hit with these things and we're not sure, and

like, how do you go about looking at it? And this is one

of those areas where I'll say this for Hiram's sake, my Batman

tool belt of things that I've got, I don't know what tool I'm

pulling out to do this forensic work around deep fakes, and is

there anything available to us?

Yeah, the shadow, it's got to be crazy on

this.

Yeah,

yes, I've got bad news. It should have been the

name of my talk, because I've got bad news. So, and schools, I

don't fault them for this, but we very often want like tech

solutions to tech problems. We want to buy a thing off the

shelf and have that be like the plug and play solution, I think.

Actually, in some ways, schools are prepared for this moment

because we saw that AI detection in the English classroom really

failed us. Yep, and so we already kind of have a mental

model for the number of false positives and stuff that we get.

The truth is that there are tools, Reality Defender is a

very well-known one. There are a couple in that space, but they

have plenty of false positives. Positives, they're not perfect

in the Washington Post. When they did their article about the

deep fakes of Nicolas Maduro, you may have seen when he was

arrested, it was accompanied with 100 different fakes of him

also being arrested. And then Washington Post did their

article about what was real and what wasn't, and they mentioned

some of the tools they were using. And I talked to a mentor

of mine, he's a guy named Perry Carpenter works for No. Before,

wrote the book "Fake. He's great, and I asked Perry, "I

said, like, what tools are they using here? And he said, "Let me

tell you a little secret about those tools. When you're the

Washington Post and you have a subscription, you're not using

just the digital tool, you're using the team of digital

forensics people, because they're the only ones who can

really do the verification.

But that's an interesting point, because,

like, you mentioned, know before, that's the service that

we use, and most of the work that we do around is phishing

training, because that's what's tied to our policies, the cyber

insurance that we have to take out, that's where that's tied,

none of this is there, and what Hiram shared with me was one of

these tools that they were using, I'd love for an

organization like no before to have that second tier,

potentially that we could leverage for them and go in on

those.

Can I do some breaking news on the podcast?

Let's do it. We're always here for that.

We're talking about that.

Yes, so already here.

So it may end up being a company like Adaptive

Security. They do cybersecurity training for like everything

from Major League Baseball to the US Senate, or it could be a

no before, but I get requests. People in the space are

interested in exactly this, so there is some awareness from

those companies that this is going to be important for

schools.

The latest lesson that I sent to my faculty was on

AI voice cloning, that very issue, and it came on the heels

of a phishing attack at our school, where one of our

students was inadvertently fished, and it just started to

spread like wildfire across our domain, so it was just an

excellent opportunity to revisit, do the PSA, which Bill

and I do frequently at our schools, but it reinforced for

me, especially after your talk, the next level is my faculty has

become quite savvy now at identifying phishing schemes, or

they send them to me. My students are not, and so I need

to take it down to that next level now. They may be able to

spot AI fakes and whatnot, but sometimes we think they know

more than they really do, and there's that next level of

education from a life skill approach, because the way we

approach it at St. Chris, and I didn't think you at MKA, it's

your professional world and your personal world, they overlap,

and we're going to teach you skills that encompasses both of

those worlds.

Yeah, it's interesting how many

cybersecurity skills have overlapped with things like

social emotional learning, like I mentioned, the sort of

emotional piece, like monitoring your own emotional response,

that's something that could be taught in that sort of self

realm, but it also has security implications and safety

implications.

Absolutely,

as a parent, creating a judgment-free culture

in the home, like learning how to listen and create that

culture where a child can come forward and know they're

supported. A lot of this is revolves around soft skills,

like the overlap between cybersecurity, child safety, and

some of these softer things that we might not associate, they

really matter.

So, yeah, what's really interesting about

this part of the conversation is we wish there was a tech

solution. We wish that there could just be a one size fits

all, gonna catch all this stuff, and what you're saying is it's a

human retraining.

Yeah, it has to be something that lives on. I worry

about a school that's, you know, oh, 2027 school year, that's

when we talked about deep fakes, and then we kind of just forgot

about it. So, it has to live in advisory programming, media

literacy, consent education, orientation, new orientation for

staff, because in many cases an attacker will attack the newest

person on your staff, because they're the person that doesn't

know how things normally operate, and Lord knows,

independent schools, we love to announce every time we hire

someone new, we tell you the name of their dog and their

hobbies, we make it

very easy to do,

and so specifically kind of new orientation, and

specifically for those departments that have access to

really sensitive data, or that can do a financial transaction

on behalf of the school, they're the ones that really need that

code word and that training,

I wrote down two things when you were speaking.

One had to do with your digital footprint, and how we've talked

to students about that, and you mentioned that from a takedown

perspective. I actually want to go in a slightly different

direction, because you also mentioned something with regard

to what schools should be thinking about and doing with

regard to the amount of materials and the types of

materials that they're putting online, and what they should be

scrubbing off, some of the metadata, some of the other

things I'd like you to address what I'll call a balance issue,

because I'll give you two quick examples. One, my oldest son was

a high school athlete, we talked a lot about him trying to get

recruited, and we have students that are in the fine arts and

doing other things, so they want to build a profile, because that

profile is what they can then point to. So the athletic

department posting the picture and talking about what's going

on, the theater department posting the picture and talking

about what's going on, that builds a social profile that

helps in terms of placement and all these other things. But from

what it sounded like you're saying, and I want to put words

in your mouth, that could be a hindrance. The other piece of it

has to do with, and I remember years ago I used to work with a

group called Ed Social Media, and we talked about how you

build your brand and how you get the SEO to drive to your sites,

and you talked about scrubbing the metadata out, which was like

a key indicator for building SEO, you talked about removing

content, which you want more content, so that it builds a

larger footprint. How do you talk to students? How do you

talk to schools about the balance between positive

exposure and the safety issues that are now part and parcel of

what we need to deal with?

Yeah, so it's an impossible situation that we've

been put in. I think if you imagine a spectrum of responses

to this, on one end being just complete business as usual, and

another end being like lock everything down post never, most

schools are going to have to figure out sort of what that

middle road looks like, and so there's probably no answer to

this that's totally satisfying, but I think looking at 678

different small things that can be done to sort of mitigate some

of the risk or to create a bit of friction for someone that

wants to attack, like, does that photo have to be in high

resolution? Does it have to have save image as automatically,

like enabled? Can more things be behind the password protected

wall? Could we focus more on impact? So, for example, if I'm

a star debater, does it have to be a photo of me debating, or

could it be a photo of my trophy, focusing more on the

impact and the work rather than just the smiling face? Even if

it is a photo of me, does it have to be my full body, or my,

you know, what I mean, or could the angles be different,

sunsetting old content, like I'd mentioned, right? So, something

from three years ago, the head of school speaking, welcoming

kids back from Covid, that might not still need to be there.

Another thing I like to do, if you're putting your voice out in

the world, consider putting a bed of music underneath. Using

AI, it's pretty easy to separate that out if someone is really

determined to get that sample, but it adds another layer of

friction for someone who wants that clean vocal sample. And so,

again, there's no one single thing that really makes

everybody safe, but I think introducing a bunch of these

little friction points and thinking more about like what

gets posted and what doesn't is important, but to your point

about like kids growing up in the world they are and like

wanting to like promote themselves, that's just a

reality that's not going anywhere soon, so that's hard to

negotiate

the analog version of this, and I think parents

made a behavioral shift was with the backpacks, putting their

children's names on their backs, and as the child be walking down

the street, anybody now knows, hey, so and so, and we're not

having to retrain our parents and our schools about the

backpack, yeah, you know, it's

a complicated backpack,

right,

it's been interesting too, because I think

so much of the focus we've had as schools has been on AI's

impact in the classroom, and largely a lot of what we're

touching on is some of the operational considerations, the

conversations you need to have with your athletic departments,

with your communications department, with your alumni

office, all of those different areas, because I've been pushing

our school to have more of those conversations, and they're

slowly coming around, but so much of that early focus was on

the impact of AI. I mean, we've had countless speakers on the

podcast about that, and that's where that focus has been, but

you're really clearly bringing to light a lot of these types of

issues as it relates to the operational aspects of these

things, that we need to take a much deeper and harder look at,

and student safety, too. A lot of AI risks,

particularly data privacy risks, are very abstract for people to

sort of make concrete and to feel real for people. It's

really hard to get people who don't care about privacy to care

about it, because it doesn't feel concrete to them. And the

other area where I think I actually have an advantage in

talking about this stuff is that a lot of AI right now feels like

we're trying to speculate about what's coming. It's, you know,

are we in this huge hype bubble or do we have AGI by 2028 Are

kids going to be at a disadvantage because they use

too much AI or not enough or not the right way? You have faculty

on all sides of this, super emotionally charged with deep

fakes, I find a lot more consensus that it's actually

okay. This is here today, I can show you headlines from last

week. The harms are very, very real. We're getting best

practices on a couple of different fronts, and there's

not a well, I'm actually pro deep fake kind of sentiment out

there in schools, and so it's just not creating the sort of

arguments and dysfunction that a lot of AI conversations have had

in schools for last few years.

It's interesting, you can kind of

like get some alignment around it.

Finally, the

only pro statement I heard was when you were

showing the adoption dog video. Oh yeah, and Eric Heilman looked

over me, he's like, 'I'll take all of that, like I don't care

whether deep pic or not the emotional toy gotta love the

dog.

Exactly, I

often say that I oversimplify Atlas'

stance on AI when I've been out and about talking to schools

that were bullish on the opportunity of AI, but really

bearish on the privacy concerns. What about you? How do you feel

about AI in general? You speak about these kind of negative

dark holes that we're dealing with and wrestling with with

deep fakes, but how do you personally. Feel about AI right

now,

so in a very nerdy way, and maybe because of that

Klingenstein background, where they have us do this exercise

where we plot out risks in schools in different quadrants,

and I think about all the time what's high likelihood, high

impact, right? That's where my focus is drawn most to some of

these deep fake risks, because they're so devastating, and

we've seen kids self-harm. The line between sort of online

harms and real-world harms is really blurry for a kid. Those

are actually integrated worlds for a kid, and so that tends to

be where I focus more, but I'm able to sort of parse that from

and separate it from how I feel about large language models day

to day. I run a small language model off my client device, and

I'm very upset about Cloud 4.7 I'm fully bought in on all the

stuff. I love building cloud skills and agentic workflows,

and so I think I have a little bit of cognitive dissonance,

maybe, about the technology, where it's pretty easy for me to

say, this is so exciting, I'll try any new tool on this end,

and on the other end, okay, these are like real things that

really matter for very vulnerable populations outside

of nursing homes and hospitals. No one deals with more

vulnerable populations than schools do, and so, especially

deep fake sexual abuse, where the laws have changed, but our

training and policies largely haven't. Every school has to

take that seriously this year. And so, I think I probably put

those things in different buckets.

So, you mentioned the laws have changed,

specifically, like I know in trying to help our students

understand not to send pornographic images, those are

now listed as classics felonies in Virginia. Is there a similar

type of statute for a deep fake,

so it gets into sort of complicated territory,

because most states have special statutes for sexting, and those

are consensually sent images, but that's still a mandatory

reporting requirement for a school, they have to report

that, even if it's consensual, because that's like CSAM for the

entire country, it is now a federal crime to publish or

threaten to publish a real or deep fake non-consensual

intimate image, and states can add on top of that, they can't

subtract from it. So, in my state of Texas, for example,

illegal to just create one of these images, even if it stays

on your phone or laptop, and that really matters, because I

get a lot of calls now, at least once a week, I get a call from

the head of school where this is happening, and very often we

know that this kid has this on their phone, they've shown it to

someone, maybe, but they haven't sent it out, and so they've kind

of tiptoed right up to the line of this legal gray area, and so

in a state like Texas, that kid has already committed a crime,

but in another state they've only maybe tiptoed up to the

edge of it, and so from a kid's perspective, it's really

important they get a message about deterrence, not a sort of

scared straight campaign, but a very clear-eyed, sober message

about this is really serious, it's not a prank or a joke. This

is often marketed to kids as a way to do a prank or a joke, and

increasingly this is coming through the form of apps that

are marketed as AI girlfriends or AI boyfriends, where kids

take a reference photo of, like, their crush from school, upload

it, it's now this avatar, and so the world of AI companionship

and deep fake abuse is kind of merging, it's that next level of

Second Life. Yeah,

but it's interesting in that one of the

things that you said is the amount of times that this is

impacting men as much as it is women. Yeah, and I can think to

this past year there are three incidents that I was aware of,

some of which I was involved in. One was with a female student,

and two were with males, and I thought that was very, very

interesting. When you were showing the interview with the

girl, I actually thought it was because one of them was related

to an incident that happened at my wife's school that made

national news as well, and it was just interesting, just the

way those things happen, the way they're covered, the way they're

handled, and the school's reaction to it, which goes back

to my forensic piece, and I think you mentioned updating

policy around this. This is, I think, one of those areas where

we need to have a deep rethink on what that looks like.

It's interesting, and this is maybe just sort of

anecdotal, because I'm only in the first few years of this

work, but I hear from girls' schools much more often than I

hear from boys' schools, with some exceptions. Browning, for

example, in New York has been fantastic, but I'm more likely

to hear from a Harpeth Hall or a Roland Park or Marlboro school,

and I can understand why, to some extent, but I think the boy

schools have work to do here. There's some stuff that lives

upstream of this. This is not an isolated problem happening in a

vacuum. The uptick in the youth to youth sex abuse category

within independent schools, it's doubled in the last 10 years,

according to Presidium. The sextortion rate doubling year

over year, that the majority of those victims are boys. The

majority of the perpetrators for deep fake sexual abuse? Are

boys, and so there's a deterrence message that has to

happen there. So one of my goals for this upcoming school year is

just to get in touch with more boys schools, because I think

they've got work to do here too. It's not just the girl schools.

There you go, you heard it here on the pod

boys schools. Let's get Evan out on the road and. Actually, that

brings me to where I'd like to kind of land this conversation.

I know you have other obligations, and you are not off

duty yet at this Atlas conference. You have more to do.

One of my final questions is, in the realm of why you're here.

We're so lucky, so fortunate to have gotten onto your schedule.

I thought that this morning's keynote was fantastic. You're

here at Atlas, talking to tech leaders. I'm curious, why you

prioritize being here with us. Why is this an audience that you

wanted to reach?

Yeah, I think technology leaders are the

natural messenger for this message, and I was actually

thinking about that as I was on stage. I thought, man, if

there's a full room, I think 500 600 people in this room, and if

just 10 of them, which is hopefully way under counting it.

Just 10 of them come back, and 10 more schools get a policy and

a plan that is 1000s of kids that will have protections that

so many other kids never had. And this very quickly transfers,

like I said, from kind of an online set of harms to physical

harm. This has impacts on a kid that are social, emotional,

academic, reputational turns their whole world upside down.

It can be life altering, and so I hope for a ripple effect out

of this room, because there's just so many smart, influential

people who will, I hope, take this message back to their

schools, and that the independent school world as a

whole is now more in a position to kind of lead the nation, I

think we're only 2% of kids in America go to independent

schools, something in that neighborhood, but we can move

faster than a lot of the public schools can that have obviously

the red tape, and it's, it's just a whole different set of

concerns, and so I think we are uniquely situated to lead on

this issue nationally and internationally, and I think

that starts in a room like this.

Evan Harris, thank you so much for joining

us, and thanks for being on the pod. We appreciate you making

the time.

Cool, thanks for having me.

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 the atlas.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.

Thank.