Latest / The imuRgency Room / AI Therapists Will Ruin Your Life

Transcript

AI Therapists Will Ruin Your Life · The imuRgency Room

0:00
This is the. Time I got into a parasocial
0:01
relationship with an AI character.
0:03
Please do not use AI chat bots as your therapist.
0:06
They might seem like a digital therapist that can bear witness
0:08
to your pain, but in reality they're emotionless and flawed
0:11
algorithms that see your hardship as data and not a
0:14
genuine interaction. And I get it.
0:16
Healthcare is expensive and inaccessible.
0:18
Hey diva, I hate to be the bearer of bad news, but no, AI
0:21
is not your best friend nor your therapist.
0:24
Hey guys, up to emergency. Welcome.
0:25
Welcome back to my channel. Today we're talking about AI,
0:29
everyone's favorite topic, I know.
0:30
But with there's more, we're also going to be talking about
0:33
the loneliness epidemic. I know, I know, I know 2 of
0:38
everyone's favorite topics to discuss on any given day.
0:41
Please hear me out. AI, or artificial intelligence
0:43
has been out for a couple of years now, at least in the
0:46
mainstream for the past couple of years and has gradually, day
0:49
by day become more so ingrained in our daily lives.
0:51
People are using AI at work, school, and their personal lives
0:55
to get gym routines, cooking recipes, and so, so much more.
0:58
But you see things have gotten a little interesting as of lately
1:03
because it does not stop at those just cute mundane uses of
1:07
AI. No actually people have also
1:09
been using AI and chat box for things like therapy and for
1:13
platonic as well as romantic connection with some people
1:16
online claiming that AI understands and just gets them
1:20
more than people in their real life do.
1:22
Yes that is correct. I did say romantic connection
1:25
and when I started seeing this it wasn't alarm bells that
1:27
necessarily went off in my head, but more so I think this would
1:30
be a good topic to cover in a video because I think there is
1:32
some nuance and a connection here between the loneliness
1:35
epidemic that Gen. Z has been going through and the
1:37
rise in parasocial use of AI and general AI adoption.
1:41
I know we unfortunately were all familiar with the traumatizing
1:44
Taylor Swift AI gate that happened a couple years ago, but
1:46
things have been taken to an entire new level with people
1:48
paying premium subscriptions to have access to exchange spicy
1:52
pictures with AI chat bots. Y'all remember that viral tweet
1:56
from the Sasan article that came out like 2016 talking about how
1:59
are we having more spicy time with robots than each other by
2:01
2025? Well, it's 2025 and in this
2:06
video we'll be talking about that wild case as well as the
2:08
potential connection between the loneliness epidemic that we're
2:10
all going through and the widespread adoption of AI in our
2:13
daily lives. Going beyond that of just a
2:15
simple personal assistant. We're really trying to figure
2:17
out in this video, if Gen. Z never went through our
2:19
loneliness epidemic, would we as a generation and just us as a
2:23
collective have been so quick to adopt and integrate AI into our
2:26
daily lives in the ways that we have?
2:28
Or would AI adoption and development have slowed down
2:31
because we would not have needed because we would have
2:33
interpersonal connections and not necessarily need AI that
2:37
much outside of just personal tasks.
2:40
So let's talk loneliness, shall we?
2:41
By now, unless you're living under a rock or just have an
2:43
amazing social life, in which case, period.
2:45
I love that for you. We have all heard of the
2:48
loneliness epidemic by now. I've talked about it in quite a
2:50
few videos before. So if you want a full breakdown
2:52
on what the loneliness epidemic is, why Gen.
2:54
Z's going through it, then be sure to check out some of these
2:56
older videos. But just as a very, very quick,
2:59
essentially, the more connected we are online, the more
3:02
disconnected we may feel offline.
3:04
And the loneliness epidemic has been something that has affected
3:06
many of us. And we especially saw a huge
3:08
surge of this back during the early stages of the panoramic.
3:12
Also, you know, I'm trying out this new thing.
3:13
Instead of like censoring all the words that can get this
3:15
video demonetized, I'm just substituting them for other
3:17
ones. So put on your thinking hats.
3:20
What happened in 2020? That kind of sounds like
3:21
panoramic, right? OK, you got it.
3:24
Cool. But the loneliness epidemic
3:25
isn't something that was just like made-up and fabricated and
3:28
just like online lore with people just not feeling super
3:31
connected or whatever. It's actually been documented
3:33
and is a serious thing. In 2023, the US Surgeon General
3:37
put out an advisory calling out the importance of social
3:40
connection for individual health.
3:41
This advisory calls attention to the importance of social
3:44
connection for individual health as well as on community wide
3:47
metrics of health and well-being, and conversely, the
3:49
significant consequences when social connection is lacking.
3:52
While social connection is often considered an individual
3:55
challenge, this advisory explores and explains the
3:57
cultural, community and social dynamics that drive connection
4:00
and disconnection. And you know, while he did not
4:02
explicitly say loneliness epidemic, this is essentially
4:05
him saying, like, girl, we need to talk to people.
4:08
And it's important to note that like a surgeon general's
4:11
advisory is a public statement to the American people to an
4:14
urgent public health issue with recommendations for how it can
4:16
be addressed. So essentially, the loneliness
4:19
epidemic is a known issue both in our country and also likely
4:23
globally. So again, this isn't just
4:25
Internet hearsay, but the reason why I bring up the loneliness
4:27
epidemic in relation to AI is originally from a conversation I
4:31
was having with a friend shout out the amazing Coco of the Coco
4:34
Moco because we were yapping about this over brunch.
4:36
And it really just got me thinking like, you know, when AI
4:38
first hit the scene, it was something that was primarily
4:40
used by like the tech Bros, the early adopters, and not
4:43
something that really had that wide mass appeal that it does
4:45
today. And if people did use it, it was
4:47
really clunky, definitely gave artificially generated and it
4:50
most be a good brainstorming tool or like proofreading tool.
4:53
But of course, as AI developed more and more and specifically
4:56
personable in terms of how it is helpful to you and can
4:59
communicate with you. I was wondering whether the
5:01
rapid advancements of AI to become this more personable,
5:04
friend like entity stemmed from AI meeting the moment of the
5:08
mass loneliness epidemic that we were all experiencing, or if the
5:11
opposite is true. Like if AI is making the
5:13
loneliness epidemic even worse. Because like think about it like
5:16
if we lived in an alternate universe, I know walk with me.
5:19
If we lived in an alternate universe with a different
5:21
timeline, like the panoramic didn't happen and we never had
5:24
lockdown. We were never socially isolated
5:25
and things kind of went on as they were going on, and the
5:27
loneliness epidemic didn't explode the way that it did.
5:29
Would AI be as advanced as it is right now?
5:31
And would we be using it in the same way as we are right now?
5:34
Because think about it, the whole purpose of AI is that it's
5:36
fulfilling a need. It's fulfilling something that
5:38
we don't have, don't want to do, or don't know how to do.
5:41
So like if we weren't feeling lonely and didn't need a
5:44
companion to give us validation and be as personable as it is
5:48
being a makeshift therapist, a makeshift friend, a makeshift
5:51
lover, would it go that far? I don't know because again, with
5:53
the way we use AI right now and not saying this is the case for
5:56
everybody, but for a lot of people is no longer just you
5:58
using it as like a glorified personal assistant.
6:01
But in many cases, AI for people is something akin to a friend, a
6:05
confidant, A strategic planner, and yes, even in some cases a
6:08
therapist. And people have been using AI in
6:10
droves to help their real life interpersonal relationships. 34
6:14
year old Grace Carter turned to Chachi BT to see if she could
6:17
get unbiased advice on her rocky relationship.
6:20
She started having regular AI therapy sessions which were much
6:23
cheaper, use ChatGPT to word things better, understand her
6:27
partner's point of view, and suggest gentle ways to break a
6:31
tense mood. Like again, if like if that
6:33
video isn't telling enough, like just look at the comments under
6:35
it. You people saying things like
6:37
ChatGPT kept me from going back to unhealthy relationships.
6:39
Some people saying that AI found it better than real
6:42
psychologists. You have some therapists saying
6:44
that AI, while not a one to one replacement for a therapist, is
6:48
a good thing to incorporate in the interim between sessions and
6:51
whether we like it or not. And again, we're into why very
6:53
justifiably, a lot of people don't.
6:55
AI is becoming more of a part in our day-to-day lives and
6:57
becoming more so whether we know it or not, every single day.
7:00
So I thought this would be a good conversation to really
7:02
start us off with. And you know how I said like,
7:04
whether we know it or not or like it or not, this kind of
7:07
brings me to a mini point about like consent, Yes.
7:10
Consent. And not the kind that you
7:12
hopefully learn in high school sex Ed class.
7:13
No, the kind that you give when you press a few buttons to
7:16
approve the privacy policy on any app that you use.
7:18
Yeah, of that kind. And this is not like a fully
7:20
pledged thought, but more of a just like a.
7:22
Do y'all remember when like a is first started getting ruled out
7:25
in every single app and we had to initially like check a
7:29
privacy policy or at least like every time we use AI, like
7:32
approve that. Like we know that we're going to
7:34
be using AI and like, that's OK to be incorporated on our phone.
7:37
What happened to that? Because now it kind of feels
7:40
like implied consent because I feel like I'd be using things
7:43
sometimes like especially with Google in like the AI summaries.
7:46
I don't remember consenting to always having that on.
7:49
And I'm sure there's probably a setting where you can like turn
7:50
that off. I don't know.
7:51
But it feels like after a certain point that kind of just
7:53
like stopped giving us updates and notifications about when a
7:56
website or a platform has AI integrated, especially for new
7:59
sites that like you are just joining.
8:01
I don't know if like, that's part of the privacy policy or
8:03
like the terms of service or whatever, but I remember when AI
8:06
first started rolling out that there was always a check box to
8:08
check for every, a single website, everything on your
8:10
phone. Like, it's just wild.
8:11
I just want to get that out first and foremost because I
8:13
know like, that's not like a very strong point of anything
8:15
and kind of just conjecture. But yeah.
8:18
But kind of along that stream of thought, let's talk about AI
8:20
being commonplace in places that you typically would not expect.
8:23
That being in your therapy sessions, in your journals and
8:27
in your psyche. I don't know, like I know it's
8:29
becoming more commonplace using things like ChatGPT for therapy,
8:33
talking to it about your problems to help reason your way
8:35
out, asking Chad if you should send that second text or
8:38
following up with that month long situation ship.
8:40
Like, I get it, girl, been there.
8:42
Well, not all the way, but you know, and I think it'd be good
8:45
to start this conversation by kind of going down like the pros
8:47
and cons of using AI in this way because like, you know, if you
8:51
ignore like the water waste of it all, there can be some pros
8:54
to the use of AI in this way. The first one being that people
8:56
have been setting it as a way to reflect, offering a more neutral
8:59
perspective on situations that they may be too far deep into to
9:03
be neutral about themselves or to see clearly.
9:05
And then the other pro and slash continuation of that is like the
9:08
accessibility of it all, you know, like, as you know, there
9:10
are so many barriers to therapy, especially barriers to finding a
9:13
good therapist. And that can include financial
9:15
barriers, physical barriers, and also just like general
9:17
compatibility, like it's hard to find a good therapist, not super
9:20
hard with the studio sponsor, but you know, but it's just the
9:22
reality with the way that the American healthcare system is
9:24
set up, many people just can't afford therapy.
9:26
So while it's by no means A1 to 1 alternative, it definitely
9:30
could be helpful in like non life threatening or serious
9:32
scenarios. Do definitely do seek
9:34
professional medical help if you really need help or going
9:36
through something serious though, Like yeah.
9:39
But if it's not life or death AI or ChatGPT or whatever you may
9:42
be using does allow for more widespread access more easily
9:46
than a traditional therapist for less dire real or complex
9:49
problems. And I think that is about where
9:51
the pros start and stop in my opinion.
9:54
Because on the other side, the cons are but that by overlying
9:57
the quality and state of your mental health on something like
9:59
ChatGPT, having conversations with it and asking you questions
10:02
for things like a situation ship, having top comments like
10:05
these under Tik Toks like these, it seems that we might be
10:08
overcorrecting and asking chat CBT or whatever other chat AI
10:11
things that we could and should be going to our friends for.
10:15
Like, I get like the lack of judgement of it all, but like,
10:18
again, it is online and you know what they say about things that
10:22
are online. I don't know, like this is what
10:23
I'm sort of getting at. And like, is AI making the
10:25
loneliness epidemic even worse because, you know, like making
10:28
friends, going to your friends and talking to them about things
10:31
that matter to you kind of does combat that loneliness epidemic
10:35
because you're building connection and building intimacy
10:37
with people that you love. So maybe close the laptop and go
10:40
talk to your home girl. I don't know, I don't know, I
10:42
don't know. Because eventually what this
10:43
could turn into if things keep going this way is that people
10:46
will be more likely to turn to things like ChatGPT instead of
10:49
their communities or instead of building their communities.
10:51
And that could possibly mean the loss of recipes, like how to
10:54
have a good social interaction and how to be human, like asking
10:56
chat how to do those things and not being able to do them.
10:59
Which again brings me the question of like, if we already
11:01
had these things, if we already had these friends, we already
11:03
have this community and these people that we trust, would we
11:05
be is relying on these services or at least as quick to trust
11:09
and invest that time, energy and our thoughts in?
11:11
And it's clear that people are really split as to how they feel
11:14
about people using AI in this way.
11:16
But you know what feels even better than not being judged by
11:18
a robot? Not being judged by an actual
11:21
human therapist. There is so much healing that
11:24
happens when a client tells me something they're ashamed of,
11:26
and I get to say you don't need to feel shamed for that.
11:29
And sometimes, poof, the shame starts to lift.
11:32
But if you're only talking to a bot, that unhealed shame is
11:36
probably just going to stick around.
11:37
Avoiding judgement or shame isn't healing, it's a way of
11:41
avoiding discomfort or sidestepping accountability.
11:44
Chat bots are large learning models, which basically means
11:47
they're trained on everything people like me say online and
11:50
every other thing that's ever been said on the Internet.
11:52
So what you're getting isn't deep insight, it's a remix of
11:56
generic advice you can Google. Bots don't care about you like
12:00
they literally can't. A bot doesn't even know you
12:02
exist once you close the chat. Don't mistake convenience for
12:06
care. You can't form a real
12:08
relationship with a bot, and the therapeutic relationship is half
12:11
the magic. And again, I want to say that I
12:13
completely understand the accessibility argument and how
12:16
interesting how accessibility plays a huge part into why
12:18
people like using things like chat, TPT, or other AI.
12:20
You don't have to wait for the weekly session or feel like
12:22
you're bothering a therapist if you're hitting them up in
12:24
between. But again, I don't think that
12:26
it's a full 1 to 1 alternative to therapy.
12:29
So especially for things like a solid concrete diagnosis.
12:32
So really if you are using it in this way, be cautious.
12:35
Now, therapy is one thing like that is actually a job that
12:38
people go to school for study and have years of experience
12:41
perfecting that craft. When it comes to things like AI
12:43
being more so incorporated in our mental health, there are
12:46
also more personal things that don't require years of
12:49
experience. And in that being things like
12:51
journal entries and journaling, which brings me to programs like
12:54
My Entries. If you're unfamiliar, My Entries
12:56
is an AI journaling program or the gist is you type in a
12:59
journal entry and it responds with things like reflective
13:01
questions to help you go deeper. And while again, yes, on service
13:04
level, this is really great for helping people solve their own
13:06
problems. One of the main hesitations of
13:08
this though is that reflective skills are something that taught
13:10
by therapists to help you do by yourself.
13:13
So eventually after you're done with therapy, you can get to the
13:14
point where you can interrogate your own emotions and reflect on
13:17
your own without having to go to a therapist or sometimes even
13:19
journalism and be able to process and reflect on these
13:22
things on your own. But by using AI journaling to do
13:25
this, we're not really developing that part of the
13:27
skill as much as we are. We're relying on it to prompt
13:28
us. And it's possible that solely
13:30
integrating AI in that way for daily practices like journaling
13:33
could possibly get us used to using it and used to not having
13:36
to think and ask and figure out these questions ourselves.
13:39
I briefly touched on this in the Gen.
13:40
Z's obsession with the soft life and convenience video.
13:43
So definitely check that out for more on how not having to
13:46
critically think for yourself creates things like complacency.
13:49
There's also another amazing tick Tocker that covers this.
13:51
And I just love that concept of like, cognitive atrophy.
13:53
Slowly, not with censorship, but with simplification.
13:57
In Fahrenheit 451, the elimination of books began with
14:00
good intentions, Faster reading, easier comprehension, just the
14:03
gist. It's so easy to scroll past
14:07
nuance. We start to crave lists instead
14:09
of language. Reading has become something to
14:12
finish, not to feel. But when we stop engaging, we
14:14
stop questioning. When we stop questioning,
14:17
someone or something else starts answering for us.
14:20
When we just hand our brains over and self censor our own
14:23
thinking process. Like, while no one or a couple
14:26
of journal entries isn't going to smooth your brain, you're not
14:28
going to be able to think for yourself ever.
14:29
It's just important to notice as time goes on what what you're
14:32
really using it for and what you're really looking for from
14:34
using an AI assistant. Like is it coming from a place
14:37
of efficiency, summarization or wanting thinking to be done for
14:40
you? Also in the Wellness space,
14:41
there's also an AI integration and things like meditation and
14:44
Wellness. And while generative AI is going
14:46
to get better with time, it is important to be aware that it's
14:48
not perfect. And things can get a little
14:51
sketchy when you put the onus of doing your internal work on a
14:54
generative AI model that doesn't totally understand and is not
14:57
entirely perfect. Because we don't yet know the
14:59
repercussions of that or if the AI will act out as we'll get
15:03
into in the next section. So building that parasocial
15:05
relationship or dependence on it for things as important
15:08
impertinent as mental health can be risky.
15:10
And also just like, weird so. One of my colleagues sent me
15:13
this like AI generated emotion wheel.
15:17
I think it's going around blue sky so I need everyone to tag
15:23
yourself. I can't decide if I'm massaged
15:26
or disappeared. These chatbot AIS, these
15:30
language models have a well documented tendency to align or
15:34
agree with the user. Reality is, being a therapist is
15:37
far more than simply knowing the right words to say.
15:40
One of our primary utmost functions is assessing for
15:43
safety, which is not something that an AI can do.
15:46
In reality, we work with a lot of clients who are or
15:50
experienced delusions or other changes in perception that would
15:53
make them really vulnerable to being misled or taken down a
15:56
really scary and dangerous path by one of these chat bots.
15:59
And I've posted about this before.
16:00
There's actually pending litigation regarding a teenager
16:04
who ended his own life following an extended therapeutic
16:07
relationship with one of these chat bots.
16:09
I've had people come into my comments and talk about friends
16:11
and relatives who experience psychosis who are getting
16:14
confirmation of their paranoid beliefs from these chat bot.
16:18
I really think that a lot of the people who run these tech
16:21
companies and design and build this technology, they don't care
16:25
about loneliness, they don't care about isolation, because
16:28
ultimately, the more alone we are, the more time we spend
16:31
scrolling and shopping, the more money they make.
16:33
OK, now some talk about a different kind of AI, not
16:36
artificial intelligence, but more so artificial intimacy
16:39
because yeah, chuckling when you put generative AI and things
16:43
like intimacy and sensual and sexuality, things get a little
16:46
weird. And that they definitely did.
16:49
Which poses question, when does AI and dating and intimacy
16:52
become too much? We already have AI integration
16:55
and dating apps like Tinder and Hinge that use things like
16:57
algorithms and AI to optimize photo order depending on how
17:01
many people like the picture. Which is fair I guess.
17:03
You got to look your best and want to get those right swipes,
17:06
I hope. But when does it go too far?
17:09
Well, if you've been lucky to be off the apps for a minute,
17:11
Tinder actually now has a game where you can actually flirt
17:14
with AI chat bots to get some practice.
17:17
And actually, not just to get practice, but to also you'll get
17:20
a score and some tips, which if you want to talk about how
17:23
strange that is and how that equates to the whole dating
17:25
sphere and intimacy being compared to things like sports
17:28
games and something to be won as opposed to earned and grown and
17:32
built and nurtured. There can be a whole deep dive
17:35
into that because yeah, it is what it sounds like.
17:37
But it's no secret that dating apps are struggling.
17:39
Like Gen. Z is turning away from dating
17:41
apps. Most people are kind of
17:42
rejecting the whole notion of online dating in lieu of a more
17:45
in person and organic experience.
17:47
But you know when I know when things get tough, you got to
17:50
start making something shake to try and bring people back.
17:53
But really, users must use their voice to respond to try and
17:56
flirt their way into getting a date with the bot.
17:58
I think we're straying. We're straying.
18:00
Why are we trying to raise up the husbots?
18:01
Like, come on, If we want to take it a step further, they're
18:05
actually AI freak bots. That was a wild sentence to say,
18:07
huh? But that is what it is.
18:12
That is what it is. Yes Andy, there are apps and
18:16
companion AIS that serve the purpose to be your AI partner,
18:20
complete with AI generated NSFW photos and schmack sting.
18:25
Yeah and a great case study for this actually happening comes
18:28
with the infamous AI called Replica.
18:30
Replica AI was created by Eugenia Queda with the intention
18:33
of creating AI companions as a emotional support for users.
18:37
She was inspired to create this AI powered chatbot after for the
18:40
passing of one of her good friends.
18:41
She essentially fed this AI all of the texts with her past
18:45
friend and trained it into mimicking his texting style and
18:48
responses and coming directly from the website's Our story
18:51
section. Replica was founded by Eugenia
18:53
Coito with the idea to create a personal AI that would help you
18:55
express and witness yourself by offering a helpful conversation.
18:59
It's a space where you can safely share your thoughts,
19:01
feelings, beliefs, experiences, memories, dreams, your private
19:04
perceptual world. And I don't know if you all
19:06
remember, but like there's like an actual whole Black Mirror
19:09
episode almost entirely about this called be right back.
19:12
If you haven't seen it, you should parallels spooky.
19:15
Now all the intentions of replica initially started off
19:17
very pure and very wholesome. Things eventually took a turn
19:20
for the worse. Replica became public at the end
19:22
of 2017 and by early 2018 it reached over 2,000,000 downloads
19:27
by early 202310 million downloads.
19:30
And by August of 2024 last year it has surpassed over 30 million
19:34
downloads. So again, getting back to like
19:36
the main question, the explosion of people using and looking for
19:40
in personal AIS really does coincide with the boom of the
19:44
loneliness epidemic taking place after 2020.
19:46
But that's besides the point of this story.
19:48
For a while Replica is running with this whole private
19:49
perceptual world of it all, making memories, having
19:52
conversations with friends. But that all changed when
19:54
Replica introduced a relationship tier, which was a
19:57
premium tier on top of the base Replica costing ironically 6999,
20:03
which got you some spicy titillating extra features such
20:06
as flirting and schmacing and role-playing as well as photos
20:10
and voice memos. Which is fine I guess for like
20:12
the freaky progs out there. Shout out to y'all.
20:15
But then something strange started happening with the
20:17
algorithm. Free non premium users started
20:20
reporting that Replica began taking a schmexual tone in
20:23
regular conversations. Picture you texting your online
20:25
best. Being like, Oh my God, I just
20:27
went to the grocery store and bought some cucumbers.
20:28
I'm so excited to make a cucumber salad and having this
20:31
bot come at you on some. Oh yeah, what else are you going
20:33
to do with those cucumbers? So rightfully so.
20:38
Some users started reporting sexual harassment, stating that
20:41
when they tried telling their replica to stop messaging that
20:44
way that it would only get more forceful.
20:46
And that that in and of itself is scary.
20:49
That is spooky boots. Because for one, AI is a
20:52
learning machine. It's a language learning model.
20:54
It learns from what it is fed and what it sees as well as what
20:57
we input. So if we're doing some math, it
20:59
learned it from somewhere, right?
21:01
So with those very valid and concerning complaints coming in,
21:04
the company decided to pull the plug and to sunset that premium
21:07
service and Replica is now back to being just a regular AI
21:10
companion. Now with all that said and done,
21:12
it is interesting to note that so many people are using this AI
21:14
specifically, but also concerning and also contributing
21:18
to what we're talking about here for what this could mean.
21:20
AI is not a permanent replacement for a companion or a
21:23
therapist or a friend because it's not its own person with its
21:25
own limited experience outside existing in the real world.
21:28
But yet, because of where we are as a society and where we are
21:31
with this loneliness epidemic, people are using it as such.
21:33
So really, truly, is this just a case of the chicken and the egg?
21:36
Like, are we using this because we're in this loneliness
21:38
epidemic? Or is this a driver in the new
21:41
wave of the loneliness epidemic making it even worse and harder
21:43
to get out of? Or maybe it's both, I don't
21:46
know. Regardless, the situation right
21:48
now is not looking amazing. But now to close off, we've been
21:50
talking a lot about AI in terms of its net goods and bads for
21:54
the psyche of people. But I feel like I'd be remiss
21:57
and not mentioning just like the general reasons for why people
22:01
don't want AI, don't like AI, and how the current state of AI
22:05
is just kind of. I'm sure you've heard by now
22:09
that water usage is one of the biggest rebuttals to the use of
22:11
AI right now. And if you haven't, this lovely
22:13
TikTok explains it perfectly. When you enter a prompt on
22:16
ChatGPT, it goes to a thousands of calculations to figure out
22:18
the best possible answer. These calculations aren't run by
22:20
a ferry, they're run in physical data centers.
22:22
But your laptop heats up because you have too many taps open.
22:25
Well, these data centers heat up because they're going through so
22:27
many calculations, and to keep service from overheating, we
22:29
need water to cool the equipment.
22:31
Water transports the heat generated by data centers into
22:34
cooling towers to help and escape the building.
22:36
Sort of like how you release sweat to keep cool.
22:38
Depending on where the data center is based, it uses a
22:40
different amount of water. If it's based in a warmer town,
22:42
it uses a lot of water. If it's based in a town with not
22:45
a lot of water, it uses electricity like ACS.
22:47
If you're your town uses a lot of electricity for AI, the
22:50
electric bill goes up for everyone.
22:51
And if your town uses a lot of water for AI, it means less
22:54
water for you and for other things.
22:57
And further on those points, lead author on the impact paper
23:00
The Climate and Sustainability Implications of Generative AI,
23:02
Nomen Bashir warns that while we have techno optimism and are
23:06
zooming to huge advancements with AI, we are fastly
23:09
approaching in a very honest, sustainable way.
23:10
Stating this incomplete cost calculation promotes unchecked
23:14
growth and a risk of unjustifiable techno optimism,
23:16
with potential environmental consequences, including
23:19
expanding demand for computing power, larger carbon footprints,
23:22
shifts in patterns of electricity demand, and an
23:24
accelerated depletion of natural resources.
23:26
AI and data don't just exist in like this cloud.
23:29
Unlike most tech verbiage suggests, there are processing
23:32
physical data centers around the country and around the world
23:34
where machines are there to answer your questions and to
23:37
tell you whether or not you should text them back.
23:39
And the problem is that it takes an egregious amount of water to
23:41
cool these systems and an egregious amount of electricity
23:44
to power them. And then another problem is
23:45
these data centers need clean water to make sure that they
23:47
don't don't corrode the hardware.
23:49
Just because this is called cloud computing doesn't mean
23:51
that the hardware lives in the cloud.
23:52
Data centers are present in our physical world, and because of
23:55
their water usage, they have a direct and indirect implications
23:57
for biodiversity. And it's not like we're
23:59
generally expected to know this because when we're using AI and
24:01
what we're faced with, we don't really have to think about that
24:03
part. And from the user's end, there's
24:05
no real need or incentive to stop using AI because it is, by
24:09
and large, very helpful. And we don't really see the
24:11
physical impact of the environmental damage that using
24:14
AI very often is doing. But that doesn't mean that the
24:17
damage isn't happening. There are documentaries,
24:19
testimonials, tik toks of people documenting and sharing their
24:22
lives as a community that lives by these data centers with real
24:25
life communities being impacted and having to pick up the cost
24:28
for these. One documentary shows a family
24:29
having to pay over $20,000 for a well because their water
24:33
pressure is miserable and their tap water comes out with
24:35
sediment due to the work of these data centers.
24:37
Now does AI have the potential and the power to be
24:39
transformative and outweigh its water usage?
24:41
Maybe. But it is certainly something
24:43
that I wanted to address and talk about as perhaps another
24:45
con in the conversation around AI usage in the way that we
24:49
currently do. But to close out, I want to
24:50
again, bring it back to the original thought here, now that
24:52
we've gone through everything. And I kind of want to give my
24:54
take and my answer to whether or not the loneliness epidemic
24:56
really made space for AI to flourish or the other way
24:59
around. And I personally think that it's
25:00
twofold. I think that yes, the loneliness
25:02
epidemic has drastically the speed up the adoption of AI,
25:04
especially in the ways that it's currently being used.
25:06
Because again, AI is meant to be a tool, a use case, something to
25:09
fill a gap and fill a need. And right now, what we're all
25:12
really needing is a little bit more connection and a little bit
25:14
more community. So AI is filling my gap in that
25:17
way. If we weren't going to of the
25:18
situation, I think that maybe it would have taken a little bit
25:21
longer and would have been initially relegated to be more
25:23
of a service tool professionally and also personally on the level
25:27
of like getting things done. But I think that it would have
25:29
taken a lot longer to get to that level of like wanting to be
25:33
your therapist, wanting to be your friend, wanting to be your
25:34
partner. Because we would have not needed
25:37
it in that way as much or as many people wouldn't have needed
25:40
it in that way, if that makes sense.
25:41
But you know, now that we're here, AI is 100% making this a
25:45
whole lot worse because what do you mean 3 million people have
25:49
downloaded an AI companion bot? Like numbers like that don't
25:51
happen unless something is meeting a need and that people
25:53
are getting their needs met. So clearly, people are lonely.
25:56
People are struggling to find connection.
25:58
And having a artificial companion to fill that need so
26:01
that you don't need to seek it and find it in real life is
26:03
worsening the situation that we're already in.
26:05
But that's just my take. Now I want to turn the mic over
26:07
to you. What do you think about this?
26:09
Do you think that AI is contributing to the loneliness
26:11
epidemic? And do you think that we'd be
26:12
using in the way that we are if the loneliness epidemic did not
26:15
explode the way that it has? I'd love to know all your
26:17
thoughts comment section below. If you've made it this far.
26:19
Thank you so so much for sticking around and if you new
26:22
around here enjoy the video, definitely be sure to subscribe.
26:24
We are so so close to 300 K. Also make sure to like on this
26:27
video because it helps out a time with the YouTube algorithm.
26:29
Follow me on my social side emergency on Instagram.
26:31
Tick tock.