Brian S. Lowery: If you could produce
a more immersive social experience,
now everybody's having
their individual social experiences.
Like now what I worry
about with AI, with VR,
with all these kind of technologies
that are expanding,
we all inhabit our own singular world.
That is more frightening to me
than like, you know,
that we all converged
in the same experience.
[Intersections]
[Brian S. Lowery: Social psychologist]
[Kylan Gibbs: Technologist]
BSL: So what makes a human a human?
(Laughter)
Kylan Gibbs: It’s one of
those questions, isn’t it?
I mean, there's like,
two ways I would look at it.
One is from my personal life
and one is from my work life.
One thing that's interesting is like,
there's been points
when I've been spending
four to five hours a day
interacting with AI.
And the interesting thing that happens
in that is the things that you notice,
when you first start interacting with it,
oh, this is really realistic.
Similarly when people first had
black and white TV and they're like,
wow, this is like real life.
But then as you get used to it,
you start to kind of realize,
the things that make it less authentic.
And I think something
that I realized with AI is
there's certain ways that we interact
that are just more spontaneous.
There's something about the predictability
of AI that teaches you
about the spontaneity of being human.
The ways they communicate,
the naturalness, the contextual awareness.
These little things that all add up.
That's on the technical side.
On the other side,
there's something of just the shared
experience of being human
that actually I think differentiates it
from other animals’ experience.
You know, you have
a traumatic moment in your life,
and then you start to resonate
with other people's.
I feel like every time I've had
something nearly catastrophic,
it opened up a new door of empathy,
and then you start to be like,
oh man, that really hurt, you know?
Or like when you cry
about something, you're like, wow.
And then you start to remember,
like this is what usually happens to me.
I start crying about something,
and then I think about all the things
that I did for my mom or my grandma
and the things that they felt.
And I feel like there's something
in that kind of like shared experience
where we have these things
that differentiate us,
we’re all different people.
But there’s something
about those common feelings
that it all kind of arises from that.
Anyway, that's one thought.
BSL: I love that answer, and I want to say
that you're not normal in that way.
Here's why I don't think you're normal.
People anthropomorphize anything.
It doesn't have to even
be that good, right?
It doesn't have to be
anywhere near as good AI
for people to treat it
like it has some human character,
people treat their cars
like they're people.
So I'm surprised that when
you interact with it a lot,
it feels less real to you.
KG: There's something about resolution.
It's like the way of seeing the world
and you kind of increase this.
It's like the same reason
you can't look at TV that's not 4K now.
And it's someone I think who worked
on early VR was saying, you know,
the interesting thing about it was
when you stepped out of it,
you're like, oh, the real world
is actually amazing.
And it's actually really hard
to recreate that in technology.
And I think the same is true for AI,
it's like maybe for some people,
when they interact with it,
the thing that they see
is some commonality.
But the thing that I always notice is,
this is very different from
the conversations I have with my parents.
Even when it says something similar,
there’s something off.
It's those little things,
that's like I think what,
over time, will add up
as people use AI more,
is they’ll start to recognize,
and I can't even point at them like,
what are those nuances, though,
that make us human?
BSL: You just know it
when you see it and you're like,
and it's missing an AI.
I mean that's also interesting
because what you just suggested
is that the more people use AI,
the less real it's going
to feel to people.
Do you think that's
what's going to happen?
KG: I mean, there's probably
another case, you know,
it's the same way as, you know,
your Instagram and Facebook feed
isn't a real conversation.
There are certainly, kids especially,
who would look at those kinds of feeds
and feel like, oh, that's a real
representation of my friends
or my favorite celebrities
or whatever I actually think,
when it's like completely --
I shouldn't say completely --
largely false.
And I do think something similar
will happen with AI,
where some people for sure
will almost be encaptured.
And they will believe
that that's the most
realistic thing that exists
and then start to compare people to that.
But I think that, you know,
if you have that degree of empathy,
you'll be like, oh,
there's something off here.
It's the same way even if you use
a Zoom call, there's something off.
It's hard to pick it up.
But like, I’m not picking up
all the signals,
and it's the very little nuances
that you probably
just subtly pick up as well.
BSL: So you don't think
that the technology
is going to advance quickly enough,
where it’ll overcome those little things
fast enough to capture all of us?
You're not worried about that?
KG: I am definitely worried about that.
Mainly because because I think
for most people it's easy, right?
So the thing about AI is it's so beholden,
at least if you think
about like, the chatbot styles,
it's so beholden to what we want.
And that's kind of
like what people, I think,
a lot of people want
in their life or they need,
is the sense of control.
And the AI gives you the sense that, like,
I can control this anthropomorphic thing.
And honestly, one of my fears
is that people get used to that.
And what does it mean when
I get used to interacting with something
that is beholden to only
my views and interests,
and then I go and interact with a human
who has their own interests?
BSL: Do you think people want control?
I think people want to be controlled.
KG: Maybe it's a form of control, though.
To be controlled
is a predictability, I guess.
BSL: Yeah, people want the world
to make sense, don't you think?
KG: Yes, yes, I think they also
want the world to be ...
There's something about, like,
preferring predictability over optimality.
So, like, I've even felt it
when you have, you know,
a mental health moment,
you have friends who have
mental health moments.
The things that I've
always seen as interesting
is your brain and your mind
prefer to stay in a state that's familiar,
even if it's worse.
So if you're in like a depressed state,
you almost would rather like stick in that
than break outside of it, right?
So there's something
about things that are familiar
rather than actually better.
And I don't know, there's a bias
towards, you know,
you almost identifying then
with those kinds of states.
BSL: Yeah, there's research on this.
One, it's called the status quo bias.
People like things that are already there.
And two, people like to have what they
believe about themselves affirmed
if they really believe them,
even if they're not positive.
So that is true.
So, like, what does that look like in AI?
(Laughter)
KG: I mean, it's definitely interesting
to me that people seem to love like,
you talk to a lot of these things
and they sound like computers
and they sound like AI,
but people love it
because it's kind of
familiar, it's controllable.
If you start to add lots of personalities
and these kinds of things,
it makes sense in context,
but I found it interesting
that as we started developing
these AI systems
that people interact with,
they all have this kind of similar voice.
And it's a very "AI voice."
You can kind of tell
that you're talking to an AI.
Maybe that’s intentional.
But there is something there, where like,
I think people have a preference
to getting what they want
from humans, from humans,
and from AI, from AI.
But that could blend,
there's already lots of, you know,
people in certain demographics
who spend a lot of time on the internet
and they start to identify,
that's their favorite form
of interacting with people.
And so I do think
that there's a reality where,
as we move into the future,
there will be people who bias
towards that for whatever reasons.
Whether it's the comfort of knowing
that someone's not judging them,
whether it's like the format
that it speaks to you with,
that will kind of bias towards
preferring those types of interactions.
But on the other hand, I always think
there’ll be a distribution of people,
and you'll have some people
who really don't like it.
And, you know, like I was saying,
the more that I interact with it now,
I find it almost painful
because I just pick up
on so many of these issues
that you're like, I can't even use it
at a certain point.
And, you know, you'd think
that, like, you know,
I’m in the AI space,
and I write 20-page docs.
I don't need AI for a single bit of it
because it does remove that voice.
And I do also wonder, though,
as people interact with it more,
will they either identify the differences
or start to conform to the things
that they're trained with AI.
It's the same as if you interact
with your partner for example, right?
You start to be biased
by the communication
because you're talking so much.
BSL: You mean they're training you?
KG: They're training you.
Your partner is probably like, you know,
they have a preferred way
of communicating.
You get used to it, these kinds of things.
So I do wonder if, as people
interact with AI more,
that they'll kind of all converge.
That's probably one
of my biggest fears actually of AI.
BSL: I'm concerned
about the exact opposite.
I'm going to shift a little bit.
So when we talk about AI,
what you're describing,
it's usually like dyadic interactions.
Like, I'm interacting
with one AI, one agent.
But really what people do is interact
with multiple people, right?
You interact in some community
or some small group setting.
And I'm surprised that there's
not more of that in AI.
So you're also in gaming.
I don't really game,
but my understanding
is that a lot of the gaming
is about connecting with the people,
and it's a community kind of experience.
So there's two things.
One, I'm really surprised
that AI seems so focused on these,
like, one-on-one interactions
as opposed to like, multiple AI agents
creating a more immersive
social experience.
KG: I love you brought it up
because that's really what we do.
BSL: Good, so that's one.
Other thing, like, the reason I worry
less about convergence
and more about divergence
is if you could produce
a more immersive social experience,
now everybody’s having their
individual social experiences.
Like now, what I worry
about with AI, with VR,
with all these kind of technologies
that are expanding,
what we can control
about our social environment,
about our physical perceptions
in the environment,
is that we all inhabit
our own singular world.
That is more frightening
to me than like, you know,
that we all converged
in the same experience.
KG: Well, my mom’s a grade-seven teacher,
and the one thing that she said
is really interesting
is if you went back like 20 years,
everybody was watching the same TV shows,
and they come to class
and they'd all be talking about it.
And now everybody watches
their own favorite YouTube channel.
And it's the siloing of reality.
Like, what we do is when we work
with games, for example,
one of the interesting things is like,
as people play through games,
it's basically the same thing.
You could have a million people
go through a game,
and it’s some differences
but you're largely going
to hit the same points.
And so one of the things
that we think about is,
what does that mean for agency?
The way we interact with media
changes the way that we feel
agency in the world.
So if we see inert media
that we can't change,
it also gives you this sense
that you can't change the world.
And so to your point,
one of the things that we want to do
with games is, how do you make it
so that each person can actually
influence that outcome?
And as you add more agents into that,
that you see, OK, I interact with this one
and it has a cascade effect.
I love it.
I mean, even in some
of the stuff we've done here,
the magic actually happens
when you do have those agents interacting,
because then you’re also not just seeing
like that one-to-one interaction
but the emergent effect
of basically that interaction.
And another thing is, if your main
controls that you have in the computer
is like point-and-click
or, in games, jump and shoot,
we're trying to see like,
what does it mean if social skills
like interaction like this,
are the ways that you actually interact
with the games, the technology
and the agents.
That’s a very different way of conversing
or of dialogue than button presses.
And I think that changes the way
that you sense agents in the world.
Because I think the way that most people
change the world is by speaking
and interacting and interacting
with other humans,
not by pressing buttons.
I mean, arguably it's the case in some.
BSL: You know, the other thing
that's interesting to me is
I don't think people have an understanding
of the systems they exist in, right?
People think about themselves as existing
in like individual relationships,
and they have a harder time
understanding system affects
like I affect you,
which affects your partner,
which affects your partner's
parents, right?
That is a harder thing to grasp.
But I think there's something
that's fundamentally human about that.
Like you are also impacted by all
these different things going on,
like, we had the person come
and put on our makeup,
and now I'm looking beautiful
and it's affecting
everybody else around me.
(Laughter)
KG: It's glowing.
BSL: Exactly.
How does that fit in?
I just haven't heard people
talk about it in that way,
which is surprising to me,
because that, I think,
is what fundamentally makes humans human.
It's interaction
and complex social situations.
KG: And these like, nested systems.
And like, they all affect
each other, right?
You think that your small activity
doesn't affect whatever
higher-level political stuff,
but it's all aggregate.
And it's all interlinking as well.
I mean, it's like the AI thing
is interesting too,
because I often hear people talk
about it as like this evolution.
You have like, you know, singular cells
to monkeys to humans to AI.
Whereas like, you could flip it,
where it's like more like, you know,
cells to organs to human to AI.
It's a system overarching
that just because it's trained on us
and we do these things, that we actually
influence that system,
now that people are interacting with it,
it has this interplay.
And that's interesting too,
when it becomes like, you know,
AI isn't this singular entity.
It is more of an institution
or a system almost
that is kind of overarching
everything else.
BSL: And it's also weird because it's
like our vision of who we are.
So when we talk about AGI,
it's like we don't even know
what intelligence is
and we think we're going to produce
something that has it.
It's just an interesting situation
where we talk about it, as you said,
it's natural evolution,
but in fact we’re creating it,
and it's not clear that we know exactly
what it is we're creating.
KG: I actually think that one
of the most interesting things
is that we're starting to work
on AI at a point where,
like, I still think we're figuring out,
you know, ourselves.
Neuroscience is very early on in its days,
and yet we're creating things
that are like, based
on our own intelligence,
and we don't really understand
even what's going on inside.
And so to your point on,
what are the effects?
We don't really know yet.
Every year a new paper comes out
and changes how people think
about child rearing.
Like how to bring up a child well,
like all those kinds of things.
And now we're creating systems
that will, you know,
kind of be overarching other humans.
What does that mean, I don't know.
I do actually think,
I happen to be in AI,
we happen to be at this point in time.
But if we could pause for a second,
I think it would be good, another
few centuries of figuring out what we are
and understanding that a little bit better
before we created something
that was in our image.
Because, we’re kind of just, you know,
it's kind of like taking a photograph
and like painting it, right?
You're not actually getting
the person and painting it.
There's something about the life
that's missing there.
So I do agree.
I think that we're honestly
kind of premature,
but I think it's just how, I guess,
you know, life goes that things
come out when they naturally should.
BSL: So, I mean, you work in AI,
so what's the most
exciting thing for you in AI?
What's your hope for it?
KG: I think it's kind of back
to that agency question.
So I mean, you know,
you read a news story,
you read a book, you watch a movie,
you watch a TV show,
this is specific to, like, my domain,
like there's something
about the communication
that we're having right now where, like,
I'm adapting to the things that you say,
to your body language,
all those kinds of things, right?
To, like, the people in the room
we have here, all these things.
And so when you have ...
AI able to, sort of, help that adaptation
so that you have that agency
in the things that you interact with.
I don't necessarily believe
in fully personalized media
because I think we need like
a shared social context.
The reason we watch a movie or a TV show
is because then we can all
talk about it, right?
But there is something about the fact
that we're all interacting
with these internet objects.
And so the way that technology feels,
you're on a screen, it doesn't change.
You're in a movie, it doesn't change.
You're watching Netflix, it doesn't change
depending on what you do.
And I think that changes the way
we see our own agency in the world.
And so I hope with AI that one
of the things that it does
is kind of opens this door to agency
in the way that we interact with media
and technology in general,
such that we do notice that effect
that you have on systems.
Because even if it's small, right,
where I take a certain action
and it completely changes an app
or it changes an experience,
maybe that helps us learn that we have
an effect in the social systems
as well that we're affecting.
So something to that effect.
BSL: So you want to make
our agency more transparent.
And do you think it does that,
because right now I'm not sure
it doesn't obfuscate our agency.
KG: No I don't necessarily know.
I agree, I mean this is why I think also
media and games is, you know,
the domain I mainly focus on.
And I think it's interesting, especially
because young people use it a lot.
And so I've heard very veteran
game developers say
how people interact with games
kind of trains kids how they
should interact with the world.
So even people who tend to be
professional players in different games
have different psychological profiles
because they bias towards certain ways
of interacting and seeing the world.
The same way, I guess, if you trained
in something academic, right,
you have a different way of viewing it.
And so if we make games and media in a way
that you feel that sort
of social impact as well,
maybe, maybe it opens the door to like,
another realm of understanding.
But, yeah, I agree that like a lot
of the systems that we have today
give you maybe a false sense
also of agency
where like we were talking
about the AI systems,
where you actually feel
like you're controlling this thing,
whereas maybe it's also biasing,
you know, and "controlling,"
having some influence over you as well.
BSL: So where do you think
things are going?
So there's obviously a huge
race among some very,
very well-resourced
organizations over AI, right?
You know, Microsoft, Google,
I mean, are the biggest maybe.
And they are very quickly
going to need to monetize it
because this is what those companies
are designed to do.
Like what do you foresee?
Because I just look
at social media as an example.
I think, at the time
when it first came out,
people were really excited,
as a new way to connect with people
and a way to stay connected to people
you know, you couldn't otherwise;
catch up with people
you lost contact with,
that sort of thing.
And it changed into something else.
In large part because
of the way it was monetized,
like going to ads, focus on attention.
What's the trajectory of AI?
KG: You know, I'm taking guesses.
BSL: Yeah, of course,
we're all taking guesses,
I won’t hold you to it, don’t worry.
KG: I think that the reality is,
we were kind of mentioning before
about the challenges of scale.
And when you invest tens
of billions of dollars in something,
you need scale.
And I think that's one of --
the way that AI is developed
and specifically even the types
of models we're using,
the economic model of it,
which is effectively
the more compute you have,
the better models you can create.
The better models you can create,
the more usage you get.
The more usage you get, the better.
So it has somewhat of a, honestly,
like monopolistic tendency, I think,
in the way that actually even
like the architectures
and the economy of it works.
And so
I think it's almost inevitable
that whatever AI systems are produced
by these large organizations
will be pushed to scale
as quickly as possible.
And there's some pluses in that
where like, you know, sure,
they're building in feedback loops,
people can give their input, it biases it.
But also at the same time,
what does it mean when a single model
is fit to a billion people, right?
So that's kind of what I meant
about the converging effect
where, what happens when we are pushed
to produce something
that fits to a billion people?
There's a lot of diversity in there.
And so, you know, we create these scaled
systems that are fitting with the whole,
like, trying to fit the whole planet.
Does that work?
And so I think what will, you know,
we're going to go through this
phase where like, yeah,
you're going to have a billion people
interacting the same AI.
And I don't know what
the effect of that will be.
Even the monetization models now
are kind of you pay
to use these kinds of things,
which are maybe OK,
but certainly ads will probably
enter the equation.
Also, what happens when you want attention
and AI is much better at that
than the algorithms
you even have on YouTube and Instagram.
And you can start
to capture that attention.
And so I certainly think it's going
to be an interesting little bit here now,
as we see these huge organizations
spending tens of billions of dollars
and the choices that they make
to then monetize that,
and what that means
for how AI proliferates.
I know a lot of the folks
in the organizations,
and their interests have
never been in that domain.
But at the same time,
you're beholden, you know,
to stock market interests and whatever
it is, then what happens?
It shifts it, right?
We're in a capitalist world.
And that's kind of like, you know,
what ultimately will change
the incentives.
So yeah it's interesting.
I mean I am interested in,
coming from your background,
you have a very different stance on it.
But, you know, it's all this
AI stuff is interesting.
But, you know, when you think,
almost to your first question,
what makes us human and like as people,
just technology in general
and specifically with AI,
like where can people find
the meaning in their life,
the values that they find true?
And how will that change, do you think,
I guess, with like the advent
of these new technologies?
Or how have you seen it change
with the technologies we've already
seen come to life?
BSL: This is going to sound
like a funny answer.
I think people are too worked up
about technology, personally.
I mean, you know,
we had this conversation.
I think, you know, people have been
using technology since we've been human.
So paper was a huge invention.
Talked about this.
The printing press, huge invention.
Computer, huge invention.
Internet, huge invention.
AI, great, another huge invention.
And through all of that, I think
what you see in a lot
of the biggest technologies
is the desire to connect
with other human beings.
I think what fundamentally makes us human
is our connection to other human beings,
our ability to engage with other human
beings, and like consciousness
and all these other things I think
are necessary preconditions.
But really, what makes us human
is connections with other humans.
And that is incredibly complex.
And I don't think we're close in terms
of technology of replicating that.
I mean, even what you described
it's like you have this feeling of like,
this isn't right, this is off.
And even if you felt like it was right,
it still would be off in ways
you didn't quite get.
I don't think we're close.
Though because it's designed to pull
our attention away from other things,
I think it impedes our ability
to do what we all kind of want to do,
which is interact with each other.
And it might change the way
we interact with each other
in a way that might feel less fulfilling.
And I think you see some of that
in social interactions now.
Some of that I mean, recently maybe,
COVID was an issue.
But, you know, people feeling less
comfortable in face-to-face interactions.
Like people dating,
there's no serendipity in hanging out
and you meet who you meet.
It's like you're using an algorithm
to try to present to you options.
That's a very different world.
So even that's prior to AI.
And I don't know how AI
is going to further influence that.
KG: And I guess just
even like the general point,
how core do you think
the need for connection is
in the sense that, you know,
I've heard some parents say
that, through COVID,
their kids went through
a major change, you know,
these regressions,
their different habits
and these kinds of things
because they weren't
connecting with people.
And then it's taken years
to overcome that.
So I do also wonder, like, you know,
whether it's through technology
or things like COVID
or just like circumstances,
could we lose that need for connection?
Or even if we need it,
you know, we might lose the desire for it
and feel emotional trauma as a result,
but still not go for it.
Like, how core do you think it is?
And do you think we're safe
in that kind of need?
BSL: So I'm going to give you
the most extreme answer,
which is I think the true one.
That you will cease to be human
if you don't have a need
for human connection.
Like I think you will be
a physical person,
but you will literally break down
as a human being.
And this is why in part --
Social isolation
or solitary confinement
is considered inhumane.
Because people literally break down,
you will start to have hallucinations.
You will break down mentally
and physically absent human connection.
So I don’t think there’s any possibility,
in my mind, of losing the need.
Like, you may get less than you need,
and that will have negative
consequences for you.
But I'm not worried about people
not wanting to be around people.
KG: Are you worried that, like,
things like social media or AI
or any of these things
could give you that sense
that you're fulfilling that need,
but not actually fulfilling it?
It's totally true, right?
Solitary confinement is a great example
because we need it.
We absolutely lose our sanity
as well as, you know, our well-being.
But maybe we can, you know,
technology can manufacture
the sense that we're fulfilling it.
And then over time,
we see these mental health crises
evolve as a result?
BSL: Yeah, that's a good question.
I think it's unlikely, but I don't know.
Honestly I don't know.
I'll talk about meaning for a second.
And I think of that as fundamentally tied
to our need for connection
to other people.
I think sometimes we confuse,
for example, our need for meaning,
with a desire for personal achievement.
That we chase personal achievement
and what we're trying
to do is generate meaning.
So I think we can be confused
and we can have those needs displaced
into less productive routes.
But I don't think it's going away.
But, you know, I don't know
that it's the healthiest.
KG: No, I'm totally aligned.
Thank you, Brian,
that was an awesome conversation.
BSL: It was great to talk to you.
Really fantastic and
super informative. Thank you.