Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:00
When you listen to Nobody Listens to
0:02
Paula Poundstone, the comedy podcast, you learn
0:04
stuff. I've been learning to throw
0:06
a boomerang, because this is the kind of
0:08
thing that really gets the listeners engaged, you know.
0:11
Interviews with people who will make
0:13
you smarter. Does the amount that
0:15
you learn protect you from cognitive
0:18
decline? Paula, don't touch
0:20
that! Can't people just listen to
0:22
the show? Can't they just enjoy a delightful
0:24
treehouse full of information and co- I think
0:26
I'm bleeding. Join us and be
0:29
a nobody. Just
1:01
as a warning, today we're talking about online safety for
1:03
kids, and with that comes discussion of the types of
1:05
abuse that can happen to children. Discretion
1:07
is advised, especially for individuals who may
1:10
find such topics distressing or triggering. So
1:15
how can we best protect kids online without changing
1:17
the entire internet? You absolutely
1:19
cannot. I think you can,
1:21
you just have to nuke the entire internet
1:23
and build it again, new. Well,
1:26
we just asked if you could do it
1:28
without changing the entire internet. Well,
1:31
you can't. Nuking it is a pretty big change.
1:34
So with our guest today, we're going to try to figure that out.
1:37
Welcome to What the Hack, a show about hackers,
1:39
scammers, and the people they go after. I'm
1:42
Adam Levin. I'm Beau Friedlander.
1:45
And I'm Travis Taylor. What
1:58
the Hack of the
2:00
safe online initiative of the global partnership
2:03
to end violence against children. Maria
2:05
Manojlovich, welcome to What the Hack.
2:09
Where are you coming to us from today? So
2:11
I'm dialing in or
2:13
streaming in from Geneva, from Switzerland,
2:15
but I'm originally from Montenegro. Switzerland.
2:18
I wish I were there right now. Many people do. They're
2:21
the worst places to be. Can you tell
2:23
us a little bit about your background and what
2:25
led you to child online safety? I
2:28
believe that so much of the trauma and the
2:30
damage of the world today really
2:32
stems from the violence of the past. As
2:35
we are seeing children affected at new scales
2:37
and speeds and through new technologies in the
2:39
online world, I think that
2:41
we risk amplifying and recreating cycles of
2:43
that abuse and violence. So
2:45
when I started working on children's rights almost
2:48
two decades ago, I quickly
2:50
realized two things. One is
2:52
that dealing with violence against children
2:54
and solving those issues is one
2:56
of the single most impactful things that
2:58
we can do because preventing
3:00
violence against children will unlock all
3:03
the other areas of their development
3:05
in life. Education, health, socializing, kills,
3:08
work, everything else. If kids are victims of
3:10
violence, they can't do any of those things.
3:13
Number two is that violence
3:16
against children is becoming increasingly
3:18
facilitated or amplified by digital
3:20
technologies. So those are
3:22
the two things that really made me compelled to
3:24
work in this space because I really believe there
3:27
is no single more important kind of a job
3:29
that I can do in the world. And
3:31
before you became an advocate for children's rights
3:34
in the digital space, what inspired you to
3:36
become an advocate for children in general? Well,
3:39
I come from Balkans and Balkans are really,
3:41
really complex. Geography
3:44
and space and growing up I witnessed
3:46
a lot of inequity, poverty,
3:49
civil unrest and wars. And
3:51
again, children suffer the most in
3:53
these situations, being able to really work
3:56
on giving the voice to the voiceless and
3:58
children are usually voiceless. was one
4:00
of the things that I always wanted to
4:02
do. So it's a very personal mission that
4:05
I have. All right. So you're from Montenegro
4:07
and you're no stranger to the
4:09
kinds of dangers that children see in
4:11
real life, not online. How
4:15
do you distinguish between online
4:17
and in real life safety
4:20
for children? I don't like
4:22
to define child online safety. I'd like
4:24
to define child safety as principled human
4:26
right and child right. And
4:29
look at the lens of how that safety
4:31
is ensured in different environments. When
4:34
you think about the online space, the same
4:36
rights that kids have in the offline world,
4:38
applied in the online world, but the settings
4:40
are somewhat different. What
4:48
are some of the biggest threats that children face online
4:50
right now? And what are some of the most common
4:52
threats they face online? Currently, I
4:54
feel there are two major
4:58
harms and risks for kids online.
5:01
One of the ones that is mostly spoken
5:03
about is cyberbullying, because I think it resonates
5:05
with a lot of people and it affects
5:08
a lot of majority of kids online.
5:10
But in addition to cyberbullying, and this
5:12
is what I'm particularly focusing on, we
5:14
are seeing that grooming of children for
5:17
sexual sexual purposes as well as extortion
5:19
are probably the most prevalent threats for
5:21
kids online. In addition to the
5:23
same ones being the most serious ones. If
5:26
you look at the reports of the National Center for
5:28
Miskarexploited Children in the US, you will
5:30
see that in 2022, 32
5:32
million of suspected reports of child
5:34
sexual abuse were made by
5:36
industry to the center. And
5:38
this is the highest number of reports in one year.
5:41
If in 1998, if you look at the figures in
5:43
1998, there were those reports for 3,000, another
5:46
32 million. In 2022 of
5:48
these 32 million reports, they contained
5:50
87 million photos and
5:53
videos of children being sexually abused.
5:56
And oftentimes people say, well, you know, 32
5:58
million is not a crime. a big
6:00
number because this contains images of kids
6:02
who were abused maybe a decade ago.
6:04
It's just the same image being recirculated.
6:07
But the fact that out of these 32
6:09
million reports, 50,000
6:12
were marked as urgent to act because a
6:14
child was in an imminent danger. That
6:16
gives you a bigger scale, a bigger picture of
6:19
how the things actually look like. And
6:22
we've been doing a lot of research on this because again,
6:25
it is not often that people ask kids
6:27
for their experiences and opinions. So we
6:29
have funded a big 25-country
6:32
scale, 25-country research called the South
6:34
End Career. And the research has
6:36
been completed now in 13 countries in southern
6:38
Asia and southeast Africa. And
6:41
what we have discovered is really
6:43
astonishing because in many of these
6:45
countries, up to 20% of kids
6:48
have experienced some form of online child sexual
6:50
exploitation and abuse in one year alone. When
6:53
you scale that to population, this
6:55
is 5 million plus kids in
6:58
2021 alone. So
7:00
then I looked for something of a comparison
7:02
for US context. And this would mean that
7:04
5 million kids is akin to every middle
7:06
and high school student in California. So
7:09
imagine every middle and high
7:11
school student in California being sexually abused
7:13
online. So there's
7:15
a huge, huge, huge pandemic of
7:18
online abuse and we are not tackling it in the
7:20
right ways. What is the
7:22
difference between regional and
7:24
regional? Is it different from
7:27
region to region? Do you see
7:29
something different in developing countries versus
7:31
somewhere like the United States or
7:33
Germany? What are we looking at? Online
7:35
CSCA is a very high-quality CSCA,
7:38
child sexual sexual abuse. It's a
7:40
very complex phenomena. But
7:42
in many ways, some of its manifestations
7:44
are really the same in the developing
7:46
world and in the developed world because
7:49
kids' lives online are somehow a
7:51
little bit different from their local
7:54
realities. And we've seen
7:56
these with research in Ghana,
7:58
in Tanzania. in the
8:01
Philippines, kids who have different
8:03
local, cultural, religious realities in
8:05
the offline world tend to behave differently in
8:07
the online because they have their global community
8:10
online. So their behaviors become much
8:12
more risk-taking, much more different than they
8:14
would behave in situations with their parents
8:16
around, at their schools, or there in
8:18
the communities, or wherever they live. So
8:21
manifestations of these particular issues for kids who
8:23
are using internet are similar. So you will
8:25
have a lot of kids who have
8:27
normalized sharing of nudes. Kids, nudes,
8:30
or so-called explicit material of kids when they
8:32
take a photo of themselves and share with
8:34
their peers of whoever they think they're sharing
8:36
with. And this has become normalized. For
8:38
example, research in the U.S. shows us that around
8:42
one in five kids aged nine to 12
8:44
think it's normal to share nudes with each
8:46
other. One
8:48
in five of these kids have shared
8:50
their own nudes already, and this is
8:53
nine to 12-year-olds. 15% of nine
8:55
to 10-year-olds share their own nudes with somebody.
8:57
In 50% of
8:59
the cases, they've shared it with somebody they don't even know. So
9:02
these things we are seeing repeatedly being
9:04
similar in many contexts. But
9:07
when you look at the abuse of
9:09
younger kids, that's different because
9:11
in places where poverty and deprivation are
9:13
higher, it is more easy to access
9:16
kids in more vulnerable situations and then
9:18
abuse them and then record an abuse
9:20
and put it online. You mentioned CSA, which
9:22
stands for child sexual abuse. How
9:25
worried should we be about the increasing usage of
9:27
AI to be generating fake images of it, but
9:30
still depicting the same level of abuse? Very
9:33
worried. I would
9:35
say particularly because there are two things.
9:37
One is that anybody
9:39
can now create child abuse
9:41
material. And especially the open-source models
9:43
are being used and tweaked in ways that
9:45
can accommodate this very easily. And we
9:47
are already seeing in the Dark Lab reports of
9:50
predators who are exchanging
9:52
images there, exchanging all the
9:54
prompts and manuals on how you can
9:56
actually create this child abuse material. They're
9:59
like... there's
10:01
not an actual person behind them, but
10:03
they're also recreating images of
10:05
known victims and creating new images of known
10:07
victims. So the faces of known victims are
10:09
being used to create new situations when you
10:12
either video the images. There
10:14
are two things that are here critical to know.
10:16
One is that once you have
10:18
your image being
10:20
shared, either it's
10:23
a fake or real image of you
10:25
in these situations, this really re-victimizes the
10:27
victims. The perpetual victimization continues
10:29
and you can't ever get out of
10:31
it. And that's why it's so important
10:33
to remove child abuse material overall from
10:35
the internet, from online world. One
10:38
of the other problems with generative AI too is
10:40
that it requires an existing data. So even if
10:42
you're creating a image
10:44
of a fictitious person, that needs
10:46
to be a composite from previous
10:48
footage, unfortunately. Correct, and a lot of
10:50
these people who have been very
10:53
passionate about collecting
10:55
child abuse material actually share millions
10:58
and millions pieces of imagery that they can
11:00
use to do that. We've
11:03
noticed also that there are, when
11:05
people go to the platforms and they try
11:07
to bring it to their attention, that
11:10
this isn't them. This is something that
11:12
was AI generated. Then there's the
11:14
whole issue of the platform saying, are you
11:16
sure you don't have any photographs of you
11:18
out there? Since you started
11:21
your advocacy work, how
11:23
have the threats changed? And
11:27
do you see a trend in any specific
11:29
direction? Yeah,
11:32
things are, this is I think one of the most dynamic,
11:35
unfortunately, I say dynamic, it sounds exciting, but unfortunately,
11:37
one of the most dynamic fields because really the
11:39
threats to kids is changing from one year to
11:41
another. So for example, three years
11:43
ago, we did not see a lot of issues
11:45
around self-generated. So kids taking photos of themselves and
11:48
that being a huge, huge explosion of that.
11:50
Now we've seen increases of 200, 600%
11:52
from year to year because
11:56
it's becoming so normalized. And then
11:58
some of these images are really consensually shared. kids
12:00
sharing, not knowing what they're doing, sharing,
12:03
dudes with themselves with, through sexual
12:05
exploration, or with their friends, or whatever it
12:07
is, not knowing that it's actually an offense to do
12:09
that. But the second thing is that
12:11
kids are getting much more easily groomed and
12:13
queers and tricked into doing those
12:16
things. Lauren.
12:26
Mike. We host a podcast
12:28
for Wired called Gadget Lab. We do. We
12:31
do. Yes, that is correct. Tell
12:33
the good people some more about it. Well, I think
12:35
the good people should definitely tune in every week because
12:37
they get to hear me roasting you. I know. All
12:40
right, no, really what Gadget Lab is,
12:43
is Mike and I tackling the biggest
12:45
questions in the world of technology. I
12:47
like to think of it as the
12:49
best of Wired's journalism, but in audio
12:52
form. We cover the big news of
12:54
the week in Techland, but we also
12:56
offer our expert analyses and opinions on
12:58
all things consumer tech, whether that's mobile
13:01
apps, hardware, startups, cryptocurrency. Mike,
13:03
what's been a recent highlight episode for
13:05
you? We did a deep dive on the
13:07
group behind the massive Okta hack. We also had
13:09
a great conversation about Web 3 and the
13:11
Metaverse. What stands out for you? Never
13:14
Metaverse you didn't like. I
13:16
really enjoyed our recent podcast about Peloton.
13:19
And recently, the legendary tech journalist, Tara Swisher,
13:21
joined us to talk all about Elon Musk
13:24
and the future of Twitter. So
13:26
I guess we should tell people how they can listen to our pod.
13:28
We release a new episode of Gadget Lab
13:30
every week, and you can listen and follow
13:32
us on Apple Podcasts, Spotify, or wherever you
13:35
pod. This
13:39
episode is brought to you by Shopify. That's
13:42
the sound of switching your business to
13:44
Shopify, the global commerce platform that
13:47
supercharges your selling. Harness the best
13:49
converting checkout and same intuitive features,
13:51
trusted apps, and powerful analytics used
13:53
by the world's leading brands. Stop
13:55
leaving sales on the table. Discover
13:57
why millions trust Shopify to build,
13:59
grow, and run their business. Sign
14:01
up today for your $1 per
14:04
month trial period at shopify.com/tech23.
14:14
Okay, so Everyone is familiar
14:17
with this idea that I can't let go of
14:19
which is that when you're online you're in a
14:21
horrible neighborhood When you are in
14:23
the real world, sometimes you can find yourself
14:25
in a horrible neighborhood and there's some basic
14:28
rules for staying safe Don't make eye contact,
14:30
you know mind your own business Keep
14:32
to yourself And the
14:35
corollary online would be you know, mind
14:37
your privacy Be careful
14:39
about what you click and don't talk to strangers
14:43
But are there specific things that people
14:45
can do specifically children can do Specifically
14:47
things we can teach children to do
14:49
online that will keep them safer than
14:52
other things I always give example
14:54
of because I have a toddler and when I think
14:56
about How did we do like
14:58
baby proofing the apartment right you
15:00
assess what the risks exist in your apartment
15:03
You're looking at your like glass vases. You're
15:05
looking at the electricity outlets and you're like,
15:07
okay So now I'm gonna be not
15:09
waiting for my child to like put a metal
15:11
bar in the outlet I'm gonna put you know
15:13
plug there. Will you fully
15:15
minimize? Risk and will
15:18
the child be super safe? No,
15:21
like she'll fall every day But
15:23
will we minimize the potential for really
15:25
serious harm Yes We will the
15:28
same goes for platforms which feature you apply
15:30
on your platform and who do you allow
15:32
to enter on your platform? Will determine
15:34
the risk the risk profile that
15:36
you have. So for example, if your
15:39
platform is matching strangers to strangers Likely
15:41
should not have kids online there, right? you
15:43
should have some age verification mechanisms or age
15:45
assurance mechanism that can keep kids outside of there because
15:48
He's being served to strangers is not a good thing
15:51
So thinking about having vulnerability
15:53
lens as you develop your platform
15:55
whether you're a startup or you exist in big platform
15:57
applying vulnerability lens and seeing how
16:01
the choices that you're making as a
16:03
platform can impact users on it is
16:06
critical. So it's all about safety by
16:08
design and really investing from upfront into
16:11
considerations like this. Unfortunately, the
16:13
culture, especially in Silicon Valley and around,
16:15
is not like that. It's more
16:17
about actually trusting safety people are being seen
16:19
as the inner devil, right? Oftentimes,
16:21
child-owned safety people, I love that
16:24
we are being called, we are the ones
16:26
who are trying to end privacy online, like stuff
16:28
like that. And it's like really
16:30
not the case, because I
16:33
believe, I'm a huge believer, I always say, but in
16:35
some ways, it can create much more risks than the
16:37
other ways. How do you realistically
16:39
monitor age restrictions? I mean, it's one thing
16:42
to say we should, but it's
16:44
another thing to say we can and will
16:46
make it happen. How do you do that?
16:48
I mean, by investing innovation in that space, like
16:50
you need to be actually, we can't just
16:53
say current state is like this, that we
16:55
can't really do anything. There are a
16:57
lot of good age assurance verification tools that can
16:59
be used. They're less accurate than
17:01
the fuzzy age around like the latest
17:03
teens. So it's difficult to actually like,
17:05
actually assess that particular point. But for
17:07
younger than 38, it's actually really accurate.
17:10
So those tools already exist. And
17:12
there are some privacy advanced ways
17:14
that it can be done, like
17:17
tokenizing, you know, third-party assurance
17:19
mechanisms where you can tokenize each person, you get
17:22
token aim of age, and then you can go
17:24
and use a token across the first. Flushes
17:26
like that are being tested right now. But
17:29
we know that platforms have so much
17:31
data that actually accurately assesses the age
17:33
of users regardless of the information you
17:35
provide to them because they know your
17:37
behavior patterns that they already already
17:39
have access to that data. They're
17:41
using that for micro-targeting, they're using that for
17:43
like, very specific, you know, add provision, everything
17:46
is about personalization nowadays, right?
17:48
So we know that that is possible. But
17:50
we also know that it's not really, you
17:53
know, it's when you look at profit,
17:55
profit interests, people will always choose profits
17:57
over. over
18:00
doing the right thing which is unfortunate.
18:03
So, I think we
18:05
need to be much
18:07
more firm and bolded in our
18:10
request to what platforms they need to do and
18:12
also potentially much more bolded in how we can legislate
18:14
some of these things. You
18:16
can invest in innovation, I understand
18:18
that. But
18:21
it's a pretty blanket phrase, I want
18:24
to invest in becoming
18:26
richer. I mean, okay, so how am
18:28
I going to do that? That
18:30
is a real nitty gritty question, it's
18:32
not an abstract question. And
18:34
I'll give you an example. A friend of mine
18:37
is a gun collector and he
18:39
has his guns in a safe. And
18:43
his kid knows
18:45
the combination to the safe. How
18:48
safe are those guns? They're not safe
18:50
at all because the child knows how to get into the
18:52
safe. Now, there are
18:55
biometric safes that
18:57
he could be using. There
18:59
are biometric tools that we use for
19:02
cybersecurity specifically for
19:04
authentication. That
19:07
said, you're opening up a whole new
19:09
can of worms when you're saying my
19:11
minor child needs
19:13
to provide a big
19:15
company or a small company, any
19:18
company at all, their biometric information.
19:20
Sorry, that's a bigger problem. So
19:22
how do you realistically, just doubling
19:24
down on this question, make
19:27
sure children aren't there when children can be
19:29
sneak children are they make it, they don't
19:31
have jobs, their whole job is being sneaky.
19:34
Listen, kids are sneaky, kids are smart.
19:37
Kids are much better at tech than we will
19:39
ever be. So they will find a ways to
19:41
go around them. Our job is
19:43
to make that experience a frictionful,
19:46
right? The more friction points you
19:48
can create to them entering some unsafe
19:50
space, that's the best way you can do.
19:53
You will never fully safe
19:55
proof any space, so there's physical digital, but your
19:57
job and our job is to make it a
19:59
better place. make that as
20:01
hard for them as possible so
20:04
that they can't actually expose themselves
20:06
to the most harmful experiences. Second
20:09
job you have is if your platform
20:11
is, and there are good examples of
20:13
this, if your platform bears a risk, and
20:16
every platform does, making sure
20:18
that you can actually enforce your terms of use,
20:20
if you have them, if that's what people subscribe
20:22
to, you should be able to guarantee that that's
20:24
what you will deliver to them. The
20:26
platform is a European platform for live streaming
20:28
called Yuba, and Yuba was very
20:31
popular during COVID, continues to be, and it's basically
20:33
just live streaming and everything. People actually love to
20:36
see each other sleep, a big thing on Yuba.
20:38
Literally, that was their biggest friend of the year, was
20:41
people watching each other sleep. So Yuba kids
20:43
realized that unless they're going to get synced
20:46
and closed over child abuse, they
20:48
have to ensure there is no nudity. So
20:51
they said, we can't aid a shirt because
20:53
there's always going to be children on our
20:55
platform, but for us to be able to
20:57
stop that is we're going to actually ban
20:59
for adults and kids nudity. So
21:01
now Yuba is still super used, not probably as much as
21:03
it would be if there was nudity on it, but
21:06
they now have filters for nudity and then
21:08
applied across the platform. And they said, our
21:10
business model depends on us being able to
21:12
actually guarantee there is no child abuse. So
21:15
they've decided something that's going to cut their
21:17
profits, still keep them be
21:19
profitable and use to some extent, but
21:21
much more safe than another platform of
21:24
similar live streaming nature. So
21:26
there was a little thing called
21:28
Chatroulette online about 10 years ago,
21:30
and it really was just people
21:32
live streaming and getting
21:34
sent, you know, Russian roulette style
21:36
to whomever was next. And
21:39
a big feature of Chatroulette was
21:41
you just be sitting there minding
21:43
your own business and a very
21:45
odd person would be appear
21:47
in front of you doing something
21:51
that kids would describe as dirty. And
21:55
I'd have no idea what happened to Chatroulette, but
21:58
it was a daily
22:02
way for people to sexually
22:04
abuse children and other people because they
22:06
were able to be nude, they were
22:09
able to do whatever they wanted nude,
22:11
and you had no control over who
22:13
appeared before you. And it still
22:15
exists. Now that's
22:17
a really extreme example of
22:19
the Wild West years of the
22:22
internet. We're still weirdly so a
22:24
decade later it still feels like
22:26
the Wild West. Now
22:29
is that just profit motive? Because to
22:31
me that's too easy. Is
22:34
it profit motive? I mean if that's the
22:37
case then shame on everybody. I would
22:39
say it's a mix of things like
22:41
definitely the crazier
22:44
and more risk-taking
22:46
or exciting or enraging. The
22:49
platform is a create
22:51
support for people enraged or like
22:54
excited or like adrenaline rush. It
22:56
will have more users. It's driven by by
22:59
increasing number of users right? So increased
23:03
number of users we're gonna get you better ad
23:05
business or better you know data. You can collect
23:07
more data you can do all of these things.
23:09
So it's driven by the need for amassing assets
23:12
whether those assets are profits or data
23:14
that you can then again monetize or or
23:16
it's getting people more engaged you can then
23:18
play more space for whichever other type of
23:20
engagement it is. So like I do
23:23
think there is there is a big
23:26
mentality of scale and magnitude
23:29
of engagement that is that is
23:31
driving a lot of the design choices. And
23:33
the more you can engage the more
23:35
you want the same same same thing
23:37
happening right? And I think
23:40
that we
23:42
don't allow that in the offline space. It's
23:44
good to have people dropping strangers, weird strangers
23:46
in like coats like in front of you like in parks
23:48
I used to see like you know growing up in what
23:50
it said but it's in the war years right? Like
23:53
I don't so people don't do that to you like I
23:56
started thinking you know what else I think I say
23:58
thinking recently about the decay of social media. And
24:00
what i think is gonna happen soon is
24:02
that we already seeing some of the platform
24:04
decaying and being abandoned like facebook for example.
24:07
What the more the case happens the last
24:09
investment continue to trust and safety the last
24:11
you protect your infrastructure. I
24:14
started thinking about is all my spaces that
24:16
is all the abandoned buildings. Like
24:18
you know i like imagine any old
24:20
like a band and amusement park. I'm
24:23
and it's creepy it's dark
24:26
in a band of buildings like. Keep it in happens
24:28
because they're like unknown people are there is like a lot of
24:30
a lot of things like you know things to follow your head
24:32
like and i think about these online
24:34
spaces is this. Disregarded a
24:38
band of neglected space is the soon become
24:40
so much more dangerous and then you will
24:42
have this trust that for movement because. How
24:45
can the being abuse the line doesn't happen all
24:47
the same platform you gonna get a kid
24:49
on the gaming platform you gonna move them to a
24:51
snapchat you gonna move it from a snapchat to what's
24:53
up you're gonna move to a whatsapp to somewhere else
24:55
because. The darker space of the lessons that you have
24:57
the land of the more you can get out of
24:59
the child right. So i don't like there's
25:01
a lot of other than to discuss
25:03
if we think about overall infrastructure the
25:06
internet and how things happen trust platforms
25:08
as well as different platform. I
25:10
think that i'm choices if
25:13
we are to think about safety
25:15
more comprehensive but i agree with you know i
25:17
can buy i don't think that currently many
25:19
of the tools that we speak about are. Really
25:22
advanced if you want them to be but there
25:24
needs to be a consideration investment in making
25:26
safety tools more accessible more privacy preserving
25:28
more focused on the safety of users
25:31
and more. Just enabling the participation of
25:33
people because it is about democracy the end.
25:36
Kids will be losing the opportunity to
25:38
participate to advance to enjoy the good
25:40
side of the digital technology as women are currently
25:42
being expelled from the channel my space
25:44
is because of the harassment and abuse they're facing.
25:47
I'm free at this for me
25:49
the issue of participation and issue of
25:52
democratic choice as much as it
25:54
is safety and protection. You know the
25:56
first thing is that part
25:58
of the structure of the internet. was that
26:00
it was built without identity verification in
26:02
mind. It was supposed to be, you
26:04
know, this wide open, free-spirited
26:07
communication platform for
26:09
everybody. What do you think
26:11
about that? This
26:13
is a great example, Adam, of
26:15
a Utopia, Utopian worldview that
26:19
is waiting for the pop. You know what the
26:21
pop sound is? It's the sound of your head
26:23
coming out of your... Anyway, the... Utopia...
26:27
Utopian people
26:29
have... always have
26:31
that blind spot. What about that?
26:34
You know, like, I think they kind of was built,
26:36
and I'm so sorry to say this again, but it
26:38
was built by a bunch of, like, white men
26:40
who were sitting in, like, you know,
26:43
some places which were, like, echo chambers of,
26:45
like, like-minded people who do not have safe
26:47
world experiences as, like, the rest of the
26:49
world. And you bring
26:52
that in and you say, like, listen, you
26:54
had a really nice idea, and in your world
26:56
this fully applies, but, like, your world is, like,
26:59
tiny, and the actual world is,
27:01
like, so much bigger and so much fuller
27:04
of other things and other choices
27:07
and thoughts and values and cultures and
27:09
religions and approaches to life. And none
27:12
of those things are necessarily bad by themselves, but
27:14
you just open... You open the
27:17
can of worms for, like, so much
27:19
bad stuff. And we are seeing
27:21
that, like, and then add on top of
27:23
that the whole boom around use
27:25
and misuse and, you know, like, free services, and then,
27:27
but the, you know, it's free, but it's actually just
27:29
your data that we want, and then, like, monetizing the
27:31
data. But it becomes... And
27:35
I also conflicted
27:37
in many ways as an individual as
27:39
well as professional. Like, I
27:41
like the anonymity bit of it. Like,
27:43
I think anonymity is really critical for
27:45
democratic processes. Like, being able to anonymously
27:47
share your thoughts and opinions is really critical. But
27:50
then, like, full of anonymity
27:52
in any way, in every way, on
27:54
every possible platform does create a
27:57
lot of risks if you can't verify who your users
27:59
are. on specific circumstances
28:01
or specific services that it becomes neglect
28:04
because no, your customer is something you
28:06
should be doing, right? It's
28:09
also a petri dish for
28:12
disinformation, misinformation. Correct. Can
28:14
I tell you something about that? I was
28:16
just recently in Kyoto for Internet Governance Forum and
28:18
I met my personal
28:20
peer, Maria Reza, who is the Nobel
28:23
Peace Prize winner. She is working a
28:25
lot on misinformation, disinformation. We
28:28
were talking about how we end
28:31
up in these fights between privacy
28:33
and safety because we are being stated
28:35
against each other because we all deal
28:38
with these downstream issues. How
28:40
do you do quantum moderation? And
28:42
then she says quantum moderation is
28:44
such a misleading way of approaching
28:47
things because she says
28:49
quantum moderation is almost imagine a river. There's
28:51
a huge river and then you take
28:54
a small cup of water in front of
28:56
the river and then you clean it and you clean
28:58
a small cup and then you put
29:00
that cup back into the river and
29:03
then the river is still super dirty and
29:05
it's full of these things and you just keep
29:07
on cleaning small bits and pieces of it but
29:09
you're not actually dealing with the very upstream issues.
29:12
If anybody tells you that privacy and safety are
29:14
not both possible, it's because they don't want to
29:16
be dealing with those issues at the same time.
29:19
So that's a great metaphor, obviously. Well, it
29:22
makes sense given that it came from a
29:24
Nobel Laureate. You
29:26
had said earlier that the death
29:30
of privacy means that these big
29:33
platforms, they know who's on the
29:35
computer. They can tell
29:37
by the things that you're searching, by the things that
29:40
you're doing. So
29:42
they have a fairly good idea. Even
29:45
if it's the kid who knows the combination
29:47
to their father's safe, they
29:50
know that that isn't
29:53
the dad looking at risk-a content
29:55
online. That's the kid because the
29:57
kid looks for slightly different things.
30:00
And with the death of
30:03
privacy, all of these things are
30:05
totally possible to gateway and say
30:07
this is a kid, this kid can't see
30:09
this, this is an adult, this person can
30:11
see it. So instead of trying to clean
30:13
the one cup of water, you start to
30:15
think about what is the behavior of the
30:17
organisms in the river and then
30:20
identify them by their behavior and stop
30:23
them from doing certain things. That seems
30:25
to me to be the answer, is
30:28
to program something that can help navigate
30:30
the river. The
30:36
river is messy and connected and you can't
30:38
create parts of the river that are child safe and
30:40
part there or not. So what if you then
30:42
identify a child in this river and you're like you
30:44
would then just play this tiny bit of a
30:46
river and then you can cube them there. How
30:49
do you secure it? I think
30:51
there's something around really
30:55
thoroughly thinking about
30:58
the design choices
31:00
and then every
31:02
river and every platform should have
31:04
mechanism to report abuse, which is
31:07
transparent, which is very clear what it is.
31:09
You should have clear sense of if you report
31:11
abuse, what will happen next? You should have a
31:13
sense of what detection technologies are they using to
31:15
detect abuse. Those
31:17
things are not worked through, they're
31:20
not actually figured out at this
31:22
point in time because many, if
31:24
you think about, at
31:26
the beginning of the podcast I mentioned 32 million reports
31:29
of child sexual abuse happening just last year
31:31
to NICMEC, to the National Center of Missing
31:33
and Exploited Children. Out of those 32 million
31:36
reports, I think around 29 million
31:38
came from meta, from
31:41
Facebook, Instagram and WhatsApp because
31:43
they do the top the most. Do
31:46
you know how many reports came from Apple last year?
31:49
I think it was 173. 170 p. Not
31:54
thousands or millions, 173 because
31:57
Apple not only does not. anything
32:00
on their cloud services anywhere,
32:02
anywhere, anywhere. They also
32:04
don't have a transparent mechanism for their
32:07
users to actually report if they encounter
32:09
abuse or child abuse material. And
32:11
that's a very, very dangerous thing to do. So
32:14
one thing is like privacy at all
32:16
costs. Great. But then at least enable users who
32:18
are using your services to be able to
32:20
report. That's a very transparent to say way of
32:22
doing it, right? It's not privacy invading at all.
32:25
But we also have, there's just
32:27
no unified and
32:30
really consolidated way of thinking about
32:32
prevention of abuse, currently on any
32:34
platform. What does legislation looks
32:36
like that moves us in the right
32:38
direction? I
32:41
think we should not be, first of
32:43
all, scared of legislation because you can
32:45
iterate on it and you can do
32:47
various things with it. I think legislation
32:49
should set standards and guiding principles towards
32:51
what we should be all be aiming
32:53
for. I always advise
32:55
for tech neutral regulation and legislation because
32:57
you don't want to be regulating every,
32:59
every five months or like every three
33:01
years because tech is changing so fast.
33:04
But regulators, so what is principled around?
33:07
Ate appropriate experiences, safety
33:09
measures, design choices, input assessments,
33:12
looking at risk profiles, things
33:15
that can be mandated by law can
33:17
look like use an
33:19
online platform. You have to do this
33:21
type of threat assessment at the beginning of your
33:24
cycle of development. Do
33:26
vulnerability lens analysis of how this can impact,
33:29
you know, engage with the users, engage with
33:31
people who can be affected. Bring
33:33
on board vulnerable communities and people
33:35
like you can mandate processes that
33:38
can result in certain technological choices
33:40
and decisions down the pathway of
33:42
development. I am much more
33:44
against legislating specific, specific technological solutions because
33:47
you will end up in a rabbit
33:49
hole of how does this work actually in
33:51
practice or like how can you apply that technology or
33:53
that tools to a specific platform. But
33:56
we are seeing good developments, as
33:58
I mentioned before, you can. to have a nice
34:01
safety bill was adopted. It has, of course, a lot
34:03
of short-aimings, but it has a lot of good stuff
34:05
as well. EU Digital
34:07
Services Act has really good way
34:09
of looking at platforms of different
34:11
size and different development stage
34:13
to mandate different ways of
34:15
activities across. We've
34:18
seen President Biden recently do the
34:20
AI executive order because
34:23
he has not seen any hopes of
34:25
Congress and we're legislating on that. We
34:27
just need to stop treating online world as something
34:29
of a unicorn. It's so different and so precious
34:31
that we can't really ever touch it because if
34:33
it touches everything, we'll crumble and innovation will die.
34:36
But that's not the case. There are good
34:38
examples of industries being regulated and safety centers
34:40
being introduced for the betterment of the actual industry
34:43
and that we can stand across the car
34:45
industry, you know, hospitality, wherever you
34:47
go, whenever you have standards of safety, airspace.
34:51
Sure, car industry was
34:53
legislated and became
34:55
safer. OSHA
34:58
legislated lots of different workplace
35:01
things to
35:03
make workers safer. These are
35:05
physical dangers. When cars were legislated, they were
35:07
legislated to not have things sticking out on
35:09
the dashboard so that when someone got in
35:11
an accident, it didn't kill them. When
35:14
certain saws were legislated to be safer,
35:16
they had guards put on them. And
35:18
these are physical dangers. Don't deal with
35:21
your children or like don't put kids in touch
35:23
on a jet roulette with somebody who's right. Like
35:26
don't let a jet roulette creep, creep
35:28
at your kid. Like finding ways to
35:30
reduce these type of clear risks that
35:32
have been documented. I mean, the Surgeon
35:34
General of the US has issued advisory
35:37
on the impact of social media, right, of
35:39
kids. And then in a lot of the
35:41
cases, again, people are constantly doubting, well,
35:44
we can't really say social media is
35:46
causing the mental health crisis in our
35:48
adolescence because like that's only correlation, that's
35:50
not causation. Like
35:53
really? I also believe that when
35:56
you talk about legislation and
35:58
OSHA and... Yes, there
36:00
are a lot of those dangers that
36:02
have been over the years
36:06
more regulated now
36:08
and people are better protected.
36:11
But that legislation also involved
36:14
individual units, whereas
36:16
where you're talking now,
36:18
you're talking about something
36:20
that has worldwide impact
36:22
simultaneously. This is
36:25
even more dangerous than anything that we can think
36:27
of to date. Well, that's
36:29
exactly right. And I think that that's
36:31
why I'm trying to get at not
36:35
an abstract notion of what the problem
36:38
is and how to fix it, but specifically, what
36:40
are you going to do? There's
36:44
the view in which we speak
36:46
about current configuration of the Internet and online
36:48
platforms as something that is a given. And
36:50
then we have to adjust our regulation, our
36:53
lives to it, not the other
36:55
way around. How to actually change the configurations
36:57
of the online platforms to meet the needs
36:59
of humans. Why are we
37:01
OK to accept this? Basically,
37:04
surveillance based data collection models are the way
37:06
that we need to live. And now we
37:08
have to find a way to make them
37:10
more safe, but we can't actually change them.
37:13
That's my question. Like, why can't we
37:15
think bigger? Why can't we think about
37:17
changing the actual why? The
37:20
reason why platforms are not kicking kids
37:22
off the platform because they drive engagement. The
37:25
kids, kids are the kids, spend the most time
37:27
online and the kids are one in three users
37:30
of the Internet are children. One in three in
37:32
many places is one in two because there is
37:34
a larger population of kids in many, many places
37:36
in the world. So when you think
37:38
about that, kicking them off the platform is
37:41
not very profitable. Using them to monetize your
37:43
product is so creating like specific products
37:45
for kids, engaging them. We remember also the attempt to
37:47
be certain to the Instagram kids at some point in
37:49
time, which was killed because there was a huge outrage
37:51
over that. They have
37:53
to stop thinking about the online world as
37:56
a beautiful set
37:58
of platforms that we. need to then build
38:01
safety measures around instead of changing some of
38:03
the core functionalities of this platform to change
38:06
those models. And those can be changed as
38:08
well because why are we allowing serving no
38:10
commercial content to kids? That should
38:12
not be allowed. You should not be able to commercialize kids'
38:14
data. That should not be allowed. That shouldn't be illegal. These
38:17
type of things, we need to just
38:20
get some boldness and start regulating because
38:22
soon it's going to be late. We are
38:24
losing a generation of kids in these completely like
38:26
wild west circumstances in
38:28
the online world. Well let's look at it this
38:30
way and that's the solution that's disguised as a
38:32
question. And if you change the
38:34
nature of surveillance, maybe you change
38:37
the online world. First,
38:47
the bad news. SAP Business
38:49
AI won't help you generate cubist versions
38:51
of your family's holiday photos. But
38:54
it will help you understand which supplier is best
38:56
to help you roll out your plant-based packaging in
38:58
Southeast Asia or identify the training your
39:00
junior project manager needs to rise up the ranks
39:03
and automate repetitive tasks while you
39:05
focus on big innovations so
39:07
you can be ready for the next opportunity. Revolutionary
39:10
technology, real-world results. That's
39:13
SAP Business AI. This
39:18
episode is brought to you by the Weather
39:20
Channel app. Did you know the app can
39:22
help you forecast more than just the weather
39:24
with allergy tracking and fluor snapping? So you
39:27
know when to stay inside and load up
39:29
on podcasts as well as air quality and
39:31
UV indexing. So you know when to get
39:34
outside load up on sunscreen and podcasts. Forecast
39:36
more of what you love with the Weather
39:38
Channel app. If
39:47
you could walk into the United... and I know
39:50
you've worked with the United Nations,
39:52
but if you could walk into the United Nations
39:54
and you're up on the podium in front of
39:56
the gathering of all the countries and
39:58
you could give them one piece
40:01
of wisdom that you
40:03
think would help make the world
40:05
more child safe. What
40:07
would you say? I would, I would say,
40:09
and I'm talking to governments,
40:11
right? I would say to governments, make
40:15
sure that your, make sure
40:18
that businesses need to think
40:20
about safety from the get-go, regulate
40:23
that one bit, make risks assessment,
40:25
obligatory part of any business model,
40:28
whether it's looking at children
40:30
or any other type of vulnerable, vulnerable
40:32
population, you need to think about the
40:35
people and having safety by design embedded
40:37
in the processes. And
40:39
that can look like different for different models, different
40:41
platforms is the most important thing that
40:43
you can do in terms of online safety is
40:46
safety by design. What
40:48
if we just said end
40:50
surveillance economies? What
40:55
if that was the ask? I would love that. But
40:57
wouldn't that solve the problem? Why
41:01
do we accept surveillance economies? There
41:03
shouldn't be surveillance economies. I
41:05
do think that surveillance economy or
41:07
whatever you want to call it, that the
41:11
internet currently is
41:14
not designed for the benefit
41:16
of humans. And I think design
41:18
is different benefit of humans and not businesses is
41:20
a critical thing and don't collect the data with
41:22
the data should not be collected and don't monetize
41:24
people's data and don't do any of those things
41:27
that are putting people in danger. But
41:29
again, dangerous practices drive engagement. And that's
41:31
why our kids are being put in
41:33
dangerous situations because dangerous and exciting adrenaline
41:35
building and not raging activities are the
41:38
ones cool. They generate the
41:40
most content. But again,
41:42
that's just me. Thank you
41:44
for that. I like that. I would actually say that at UN. I
41:47
think you should say it. And I think that we're just
41:49
going to say, look, how about this? Let's
41:52
change everything. Say
41:54
you're just walking out of a dress in the UN and now
41:56
you're actually just going to talk to a parent. What
41:58
advice do you have for parents? keep their children
42:00
safe online. That's
42:02
a very good one. So the first most
42:04
important thing is safety by design or end
42:06
miscellaneous in quantity, right? The second one is
42:10
start sharing these conversations with your child as
42:12
early as possible. And don't assume the child
42:14
would not share their news, because they will.
42:17
So just assume, you can just tell your child, once you
42:19
decide to share your news, can you please talk to me
42:21
so we can decide about the safe way to do it,
42:23
because they will do it. Start normalizing
42:26
these conversations as early as possible.
42:29
It's just a regular parenting advice.
42:32
The more you're open to them, and the more and
42:34
the less stigma exists around potential
42:37
harms that can happen to them
42:39
online, the more likely they will be
42:41
to tell you about bad
42:43
experiences. And this is something that
42:45
we have seen really be critical for prevention of
42:48
suicide at the end of the tunnel,
42:50
kids not reporting, living with trauma, getting
42:52
off school, doing all sorts of bad
42:54
things, because they feel so
42:56
much shame and so much unbearable
42:58
sense of failure that they can't
43:01
actually cope with it. So the more you can be
43:03
open. And so my advice is
43:05
never surveil your child. My advice is never
43:07
be controlling. My advice is always just be
43:09
open to whatever the child is doing. But
43:12
be open in a way that they can come to you, because
43:14
the people think about reporting of these crimes,
43:17
there are many ways of disclosing. Reporting is
43:19
not only picking a phone and reporting. Reporting
43:21
is talking to a friend. So really build
43:23
a nurturing environment that are around your child so they
43:25
can be free to tell you about
43:27
their experiences. Which this brings us full
43:30
circle to the whole purpose of this
43:32
program, which is to
43:34
create a shame-free zone where
43:36
people feel it's OK to
43:39
talk about things that have happened to them.
43:41
There's nothing that's wrong. It's
43:43
all about sharing experiences,
43:45
sharing threats, sharing solutions, but
43:48
creating an environment where people
43:50
feel safe. And that's
43:52
what you're talking about in a family. Creating
43:55
an environment where kids feel safe, that
43:57
they can talk to you about whatever
43:59
they're thinking whatever they're doing and not feel
44:01
like they're going to get a 22-hour lecture or
44:03
you're going to send
44:05
them to their room and starve them for the next
44:07
three months. Yeah, deprive them of the ether.
44:10
And Adam, that's a really perfect,
44:12
perfect way of saying it. And I
44:14
do want us to say one more thing because
44:16
there's a huge misconception around who is the danger
44:18
for kids. There's a whole
44:20
narrative from ages ago, stranger danger,
44:23
right? That is so
44:25
false because even in the all-night world, perpetrators
44:27
are usually people that kids know. I
44:31
really learned something today. Thank you so much. Maria,
44:35
you were totally awesome. And
44:37
if people want to find out more about you and the work
44:40
you do, where would they go? Just
44:42
go check us out on safeonline.global.
44:45
Safeonline.global? Dot global. Right.
44:48
Great. Now
44:59
it is time for the tinfoil swan.
45:02
Our paranoid takeaway to keep you safe on
45:04
and offline. Well, the holiday
45:06
season has officially been here since, I don't know,
45:08
Labor Day? As far as some retailers I know
45:10
are concerned, I think it started in late August.
45:14
Yeah, not even Labor Day. Well,
45:16
listen, there's a bunch of scams that
45:18
are a lot harder to spot this
45:20
time of year. They're
45:22
not new scams, but like I
45:24
said, harder to spot. Yeah, this time of
45:27
year you get a lot more unsolicited communication,
45:29
especially about online orders and deliveries. Let's
45:31
do a lightning round. What
45:33
are the, well, this
45:36
is your term, Adam, pantheon of ishings. Go
45:38
ahead. A classic is phishing, and that's where
45:41
you get an email from someone, and it's
45:43
claiming an update for an order that you
45:45
made. But what you don't realize is you
45:47
click on the link, and it's
45:49
either a malicious attachment or it's a phony
45:52
website. As with any email,
45:54
as we continue to tell everyone, check
45:56
the sender and the URL, make sure
45:59
that it's a phishing. there's
48:00
going to be squishing which is where someone says
48:02
you fuck you or come anyway
48:05
thank you thank
48:07
you to death and then you're dead just
48:09
said I do yeah yeah most of these
48:11
will follow similar beats to one another but
48:14
the lessons are all the same you
48:16
got to treat any incoming communication
48:19
with suspicion don't respond in a
48:21
panic especially if it's
48:23
marked urgent and if
48:25
you order something over retail website go
48:28
to the source if you don't want remorse go to directly
48:30
to that site rather than clicking
48:32
on links or downloading attachments please
48:35
okay so if you don't want
48:37
squishing act with suspicion and keep
48:40
an eye on your accounts if you're
48:42
buying gifts for people are making travel
48:45
plans there's gonna be a lot of
48:47
account activity out there especially with money
48:49
coming out of your tracker
48:52
what you spend see if there's
48:54
anything that looks suspicious and just
48:56
go about your life and
48:59
that's our tid foil swan you
49:01
guys have to love squishing and we think of that what
49:17
the heck with Adam Levin is a production of
49:19
loud dream media you can find us online at
49:21
Adam Levin calm and on Instagram Twitter and Facebook
49:23
at Adam K Levin loud
49:34
dream
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More