0:01

Hi listeners, Dave here. We'll be

0:04

back with a whole new season of

0:06

how God works, starting in early March.

0:08

But in the meantime, we wanted to

0:10

share a few shows from our archives

0:12

that speak to some of the most

0:14

pressing issues we're all facing today.

0:16

One of the biggest involves

0:18

technology. AI in particular. Just

0:20

this week, the new administration

0:22

announced a plan to invest

0:24

$500 billion in AI development

0:26

over the next few years.

0:28

That's a hefty sum. But you

0:31

might wonder, what does that have to

0:33

do with religion? Turns out, more

0:35

than you might think. The

0:37

links between technology and religion

0:40

are back in the news.

0:42

In fact, a recent New

0:44

York Times article profiled a

0:46

surge of new faith-based tech

0:48

startups and new ways in

0:51

which religious leaders and institutions

0:53

are experimenting with technology, including

0:55

AI alter egos like rabbi

0:57

bot and sermons generated entirely

0:59

by AI. which ultimately raised

1:02

the questions, can God speak

1:04

through AI? Or is AI like God?

1:06

The ethics and surprising

1:09

history of Tekken

1:11

religion is something

1:13

we explored in

1:16

a previous episode,

1:18

and we're excited to

1:20

share it with you again

1:22

now, given the times. We

1:24

hope you enjoy it. I

1:27

mean, if you can count sermons

1:29

by Zoom. But in reality, technology

1:31

has been creeping into

1:34

religions for longer than

1:36

you might think. In 2007,

1:38

a humanoid robot called Mindar

1:40

began offering Buddhist wisdom in

1:43

a 400-year-old temple in Japan.

1:45

And a robot named Santo?

1:47

is helping Catholics in Poland pray during the

1:49

pandemic. It is easier for a camel to

1:52

pass through the eye of a needle than

1:54

for someone rich to enter the kingdom of

1:56

God. I know, I know, this all might

1:58

sound a little... sci-fi... But it

2:00

is year 2022. Tech

2:04

isn't just about robots

2:06

and algorithms that assemble your

2:08

car and tell you

2:10

what groceries you need. Robots

2:12

aren't just about efficiency

2:15

and utility anymore. They're about

2:17

relationships, trust, even emotion.

2:19

So now the question is,

2:21

can we build a

2:23

robot that actually seems alive?

2:33

It's been surprising to me how

2:35

willing and able people are to

2:37

engage them as social entities. And

2:40

I think it's happening on both

2:42

conscious but also subconscious levels. This

2:45

is religion's new frontier. If

2:47

you can believe, trust, or be

2:49

comforted by a robot, can

2:51

you worship with it? Or even

2:53

perhaps worship the technology itself?

2:55

Why can't God give a robot

2:58

a soul? I don't think

3:00

the limits of your imagination are

3:02

necessarily the limits of your

3:04

gods. On this

3:06

episode, we'll talk to MIT

3:08

professor Cynthia Brasile about social robotics

3:10

and a friendly little robot

3:13

named Nexie. She was probably about

3:15

four feet tall, two eyes,

3:17

eyelids, eyebrows, you know, the whole

3:19

gamut. And we'll ask religion

3:21

professor Robert Girassi what these advances

3:23

mean for faith today. Maybe

3:25

we'll have AI that evokes feelings

3:27

of awe and can raise

3:29

people to a more spiritual kind

3:31

of level. I'm

3:33

Dave Disteno and this is How God

3:36

Works. Cynthia

3:39

Brasile is one of the world's

3:41

leading social roboticists. And like me,

3:44

her interest in robots began in

3:46

a galaxy far, far away. I

3:50

have to say I'm a Star

3:52

Wars child. Me too. In

3:58

many ways, kind

4:00

of... crystallized my idea of

4:02

like what what an autonomous robot

4:04

should be socially emotionally intelligent

4:07

entities who have rich personalities who

4:09

are really there to be more

4:11

allies to us rather than just

4:13

tools that do what we say.

4:16

Of course in Star Wars they

4:18

go off and save the galaxy

4:20

together so like there was a

4:22

real value add to that as well.

4:24

So she started designing robots that

4:26

like R2D2 and C3PO.

4:29

appear to have personalities, thoughts, and

4:31

feelings all their own. In fact, if

4:33

you walk into her lab at MIT, you

4:35

kind of feel like you're in a

4:37

Star Wars movie. You're surrounded by

4:39

robots of all kinds. There are cute

4:41

and furry ones that look like teddy

4:44

bears and dragons that she uses to

4:46

help children learn. There are humanoid

4:48

ones with wires hanging out that

4:50

look like a cross between C3PO

4:52

and Terminator. There are card dashboards with

4:54

little robotic screens that pop up

4:57

to chat with you. get ready to

4:59

see those in your vehicles in a few

5:01

years. Some of these robots look like

5:03

they could actually be alive, and others

5:05

are very clearly a bunch of gears

5:07

and screens. But Cynthia says that when

5:09

it comes to people treating them like,

5:11

well, a friend, it doesn't actually matter

5:14

that much if they look human.

5:16

They clearly are designed explicitly to

5:18

look like robots, or maybe they're kind

5:20

of playful, like kind of a sesame

5:23

street character sort of brought to life

5:25

in a robotic form, but they're clearly

5:27

not human. It's been surprising to

5:29

me how willing and able people

5:31

are to engage them as social

5:34

entities, and I think it's happening

5:36

on both conscious but also subconscious

5:38

levels. You might be thinking,

5:40

not me. I wouldn't respond to

5:43

a robot like that. I would know it was

5:45

just a machine. But did you

5:47

cry when the robot Wally looked

5:49

sad in that Pixar movies? Or

5:51

did you laugh when the flying

5:54

carpet in Aladdin wildly gesticulated? Cynthia

5:57

says it's not so different.

6:00

Robots are like mechanical cartoons, you know, in

6:02

many ways I say it really does not

6:04

have to look human at all, but there's

6:06

qualities of movements that make you want to

6:08

interpret its behavior as a social entity. The

6:11

thing that's fascinating to me is even

6:13

the lowest fidelity mechanical thing seemed to be

6:15

enough where people do want to anthropomorphize

6:17

them. I remember watching it,

6:19

the magic carpet in Aladdin, and thinking,

6:22

it's annoyed. It's happy. It's excited, but

6:24

you know, I think about some of

6:26

the cute robots you've designed to work

6:28

with kids and it looks like they're

6:30

breathing. They have kind of a respiration

6:32

mode in them or the gaze and

6:34

the eyes change. And I think we

6:36

really want to see intent

6:39

in things. And the easier you make that

6:41

with that design, it grasps us. You

6:43

know, once you start dealing with something

6:45

like speech, it's like game over. You know,

6:47

it can be just text and that's enough for

6:49

people to treat a computer as a social entity. There's

6:52

an implicit assumption that the more

6:54

human, the better. And

6:56

a lot of my work is

6:58

basically saying, no, not necessarily.

7:00

You know, it's like, what is

7:02

the interaction you're trying to

7:04

achieve? And in a world where

7:07

we've got plenty of people,

7:09

why do we need robots to

7:11

try to pretend they're people?

7:13

Why not design robots to

7:15

be what they are, to

7:17

be great partners and collaborators

7:19

to compliment us and our

7:22

human abilities? It's

7:24

not about looking more human. It's

7:26

about acting more human. And

7:28

that's where cues come in. The

7:31

nonverbal things that show what we're thinking

7:33

about and what we intend to do. We

7:36

evolved to be social beings,

7:38

to collaborate, to interact with one

7:40

another. And so when another

7:42

entity exhibits these cues, I think

7:44

our brain at a deep

7:46

level is like, oh, social entity,

7:48

you know, not a thing

7:50

or object that's governed by the

7:52

laws of physics, but some

7:54

kind of being now that is

7:56

governed by the laws of

7:58

having a mind, right, of social

8:00

interaction, of having goals and

8:02

intents. Gaze is a profoundly important

8:04

social cue, right? It's attending

8:06

to you as - attention to me. Across all

8:08

species, gaze is extremely salient. So even a camera, a raw

8:10

camera, if it turns just to look at you, people associate

8:12

that with gaze. You know, is it a friendly thing or

8:14

is it potentially a threat, right? Okay, I understand

8:17

if some people might be a

8:19

little skeptical about this.

8:21

How could people really think a

8:23

robot has feelings? That it

8:25

actually wants to be nice or mean?

8:28

Well... It's not a futuristic

8:30

thing. We have the technology

8:32

right now to do this, to

8:34

build robots that can make us

8:36

trust them, or make us think they're

8:39

shifty. And that doesn't mean

8:41

people are suckers or

8:44

gullible. Responding this way to

8:46

a robot is actually what

8:48

makes us human. Don't believe

8:50

me? Well, let me tell

8:52

you about an experiment Cynthia

8:54

and I did using a

8:56

robot called Nexi. So Nexte kind

8:59

of had very much an anthropomorphic,

9:01

more human-like face. Two eyes, eyelids,

9:03

eyebrows, a jaw that could move

9:05

in ways that could suggest different

9:07

kind of jaw, mouth movements to

9:09

express. She's smaller than a person. So

9:11

like she was probably about, I think, maybe

9:14

four feet tall, kind of from floor to

9:16

top of the head. So it's about the

9:18

same height that if you're going to sit

9:20

in a chair. Nexte could be making

9:22

like eye-to-eye contact with you. had two

9:25

arms, fingers, you know, the whole gamut.

9:27

So it was essentially like an upper

9:29

torso humanoid robot, but it was on

9:32

a wheeled base. Okay, I have to admit,

9:34

Nexty was the coolest robot I had

9:36

ever seen at the time. I mean, I'm

9:38

a total sci-fi geek. But that wasn't

9:41

the only reason I wanted to work

9:43

with Cynthia. I actually wanted to

9:45

use this robot to learn something

9:47

about humans. The goal of the

9:49

experiment Cynthia and I designed was to

9:52

see if people would unconsciously react to

9:54

the robots' nonverbal cues. You know, things

9:56

like, fidgeting, touching its face, opening its

9:58

arms in a welcoming... or crossing them

10:01

in a blocking one, in the

10:03

same way that we respond to

10:05

a human's. I can communicate in

10:07

many of the ways that

10:09

people do. I can tell you that

10:11

I'm sad, mad, confused, excited,

10:14

or even bored just by

10:16

moving my face. But I hope you

10:18

can see that I am very happy

10:20

to have met you. Thank you for

10:23

visiting me, and I hope to see

10:25

you again soon. Before

10:29

we ever brought Nexsey into the picture,

10:31

we had filmed many pairs of people

10:34

talking with each other, and then playing

10:36

a game where they could either cooperate

10:38

with their partner and make a little

10:40

money, or cheat their partner and make

10:42

a lot. To make a long story

10:45

short, we found that certain cues help

10:47

people predict their partner's plans to be

10:49

fair or to cheat. That's when Nexsey

10:51

came in. We wanted to see if

10:53

people would respond to the same cues if

10:55

they came from a robot. Of

10:57

course, this was 15 years ago,

10:59

and the technology at the time

11:02

wasn't up to letting Nexsey run

11:04

all on her own. So when

11:06

people were chatting with her before

11:08

playing the trust game, Nexsey's brain

11:11

was really being run by three,

11:13

well, wizards. Not the Gandoff

11:15

type. Think Wizard of Oz. It's

11:17

really being puppeted by a human

11:19

behind the scenes. So in the

11:21

case of Nexsey, we... we actually

11:24

had a team of graduate students.

11:26

One of the graduate students was

11:28

kind of controlling Nexe's gestures and

11:30

the other graduate student was literally

11:32

having the conversation. So you could

11:35

feel like you're actually having this

11:37

kind of chit-chat interaction with this

11:39

robot. Yeah, the illusion was amazing.

11:41

And the thing that was fascinating to

11:44

me, I don't remember if I ever told you

11:46

this, but during the conversation where people were

11:48

kind of doing to get to know

11:50

you, they started next to you like really.

11:52

Interesting questions

11:55

like, nexty, what is the meaning

11:57

of life? Nexty, is there a

11:59

God? It was very, it was

12:01

very fascinated, because even at that

12:03

point, people were willing to ask

12:06

the robot in a way that

12:08

they were actually not, they weren't

12:10

trying to mess with it. They

12:12

were actually seeking those kinds of

12:14

existential questions from it. It was, it

12:16

was fascinating to me. The first few

12:18

minutes, people were like, holy cow, I'm

12:21

talking with a robot. But they really

12:23

did quickly feel comfortable enough

12:25

to offer all kinds of

12:28

self-disclosures. how they spent their summer

12:30

vacation, what their life goals were,

12:32

you know, the kind of stuff

12:34

you tell someone you were comfortable

12:36

chatting with. But the main point

12:39

wasn't what Nexie was saying, it's

12:41

what she was doing. So during

12:43

the chit-chat we had Nexie show one

12:45

of the two sets of cues, a

12:47

sneaky set or a more neutral one.

12:49

The sneaky cues were the ones we

12:51

found people showed when they were going

12:54

to cheat you. They'd fidget with

12:56

their hands, repeatedly touch their faces,

12:58

and cross their arms to kind

13:00

of signal a social distance

13:02

or block. The neutral ones were

13:05

more innocuous, like raising an arm

13:07

or tilting the head. And then it was

13:09

time to put some money on the line.

13:11

We had people play a game where they

13:13

had to decide if they thought Nexsey

13:16

was going to cheat and try and

13:18

take money from them. What we found

13:20

was nothing short of startling.

13:22

People who saw Nexsey give the

13:24

untrustworthy cues reported believing that the robot

13:27

was going to cheat them in the

13:29

game. And so, they tried to cheat

13:31

her back when it was their turn.

13:33

Those who saw the other cues thought

13:36

Nexsey was great and that they could

13:38

count on her. They had no idea why.

13:40

They just said they had a gut

13:42

feeling. The real reason, though, was that

13:44

we had cracked the mind's unconscious

13:46

code for reading whether someone cares

13:49

about you or not. And best

13:51

of all... We showed we could

13:53

use it perfectly in a

13:55

robot. It learned so much from

13:58

that project. And one

14:00

of the things that really fascinated

14:03

me was, we used Nexie

14:05

as a scientific instrument, right? So

14:07

it was acknowledging that it's

14:09

literally impossible to ask a

14:11

human to control their nonverbal

14:13

cues with such precision, but

14:16

with the robot, you can

14:18

very precisely control these cues.

14:20

And that's great, right? You can

14:22

make sure that if you're designing

14:24

a robotic partner, it can make

14:27

people feel comfortable. Like... They

14:29

can rely on it. Tell it

14:31

anything. Trust it. Well, maybe.

14:33

But it depends on if

14:36

that trust is well placed. Or

14:38

put another way, it depends

14:40

on who designed that robot,

14:43

and for what purpose. Here's

14:45

where text-rolling religion

14:48

comes into question.

14:50

If we can make a robot

14:52

people will trust, we'll

14:54

believe. That raises a host

14:56

of questions when we're talking

14:58

about things like morality, sin,

15:01

and existential concerns. I

15:03

think when you're talking about

15:05

something that's so deeply human,

15:07

like spirituality, I think

15:10

you really do have to think

15:12

about what's the appropriate role of

15:14

a technology in that scenario?

15:16

Just because we can make

15:18

a technology seem more trustworthy,

15:20

more wise, more caring, does that

15:22

mean we should? For

15:26

thousands of years people have

15:28

looked to priests, rabbis, and

15:30

imams for spiritual guidance.

15:33

Is it okay for robots or

15:35

other tech to fill those roles?

15:37

Or put another way, can tech

15:39

have a spark of the spiritual

15:41

in it? I

15:53

think people finding ways to see

15:55

magic and enchantment and wonder in

15:57

the world is a good thing.

16:00

That's Robert Girassi, a professor of

16:02

religion at Manhattan College, who has

16:04

spent a good deal of his career exploring

16:06

how technology and religion interact. And

16:09

one thing he likes to point out

16:11

is that the question of religion using technology

16:13

is nothing new. There's

16:15

a long history, for example, in Christianity

16:17

of believing technology was kind of God's

16:19

way of making the world better that

16:21

goes back at least to France. It's

16:23

very clear in Francis Bacon. Bacon

16:26

was a century philosopher and

16:28

devout Christian. Who, as Robert says,

16:30

believed God gave us the smart

16:32

to develop technology as a gift

16:34

to improve the world. And

16:36

you even have people in

16:38

early 20th century Russia arguing

16:40

that the resurrection promised by

16:42

Jesus was a resurrection that

16:44

human beings were actively supposed

16:46

to create through technology. So

16:48

there are people within religious

16:50

communities who have had that

16:52

kind of commitment to technology. So

16:55

when it comes to religion, Robert, like

16:58

Cynthia, thinks the main question for

17:00

tech isn't if it will be used, but

17:03

why it should be used, either

17:05

in current religions or to

17:07

develop entirely new ones. So

17:09

what is our motivation for

17:12

building these kinds of new

17:14

religions really matters to me

17:16

because religion engages contemporary

17:18

technology. You can

17:20

imagine how new architectural and engineering

17:22

techniques produce new kinds of

17:24

church experiences for medieval Europeans, right?

17:27

So there's kind of nothing

17:29

new about thinking about how does

17:31

this technology change the way

17:33

I'm going to do religion and

17:35

seeing divine activity through AI

17:37

or robotics seems like that could

17:40

be reasonable to me. That's

17:43

true. You know, I never thought

17:45

about that, but when you walk into

17:47

medieval churches and you have those

17:49

beautiful high arches, they evoke this feeling

17:51

of awe in you. And as

17:53

a as a psychologist who studies emotion,

17:55

we know how those feelings of

17:57

awe affect people and can actually lead

17:59

to spirit. experiences. So you're right,

18:01

that in and of itself

18:03

was a technology to kind

18:05

of ping parts of our minds. Yeah,

18:07

and maybe we'll have AI that evokes

18:10

similar feelings of awe and can raise

18:12

people to a more spiritual

18:14

kind of level. Robots that

18:16

talk to us are one thing, but

18:19

an artificial intelligence that

18:21

could truly think would be another.

18:23

You could ask it anything, and it

18:25

would know. It would seem wise, all

18:27

knowledgeable. Perhaps even godlike?

18:29

Worthy of following? Or maybe

18:32

even worshipping? This might seem a

18:34

bit strange at first. An idea

18:36

that only a modern tech guru

18:38

might come up with. But believe

18:40

it or not, the idea has a basis

18:42

in some traditional faiths. He

18:44

might be familiar with the idea

18:47

of an avatar from that

18:49

James Cameron flick, where it

18:51

referred to a remote-controlled, genetically

18:54

engineered body sent to the

18:56

people of Pandora. Many

18:58

religions have the idea

19:00

of avatars too. Physical forms that

19:03

gods use when coming to

19:05

earth to talk to us. In

19:07

Hinduism or Buddhism, they're pretty

19:09

common. So at a time like

19:12

ours, where AIs are about to

19:14

become sentient or self-aware, could

19:16

a God inhabit a computer

19:19

or manifest in the internet?

19:21

Yeah, there's so much there. I've

19:23

already seen... in some of my

19:25

research in India, people referring to

19:27

the 10th avatar of Vishnu, Kalki,

19:29

coming as an AI. So for

19:31

those who don't know much about

19:33

Hinduism, Vishnu is a preserver god

19:35

and he has shown up on

19:37

earth so far in nine different

19:39

forms and he's supposed to come

19:41

a tenth time as this god

19:43

Kalki to end the world and

19:45

begin the new world to restore

19:47

the world to its original state

19:49

of grace right because that's a

19:51

cyclical process in Hinduism. Once I was

19:53

given a talk at a design institute

19:56

and I had a student who said

19:58

to me what is God today,

20:00

but it could be AI.

20:02

Couldn't Kalki come as

20:04

AI? Wouldn't that make sense?

20:06

And I've had other interview conversationalists

20:08

who have said, hey, there are lots

20:10

of people who think that maybe

20:12

Kalki will come as AI because if

20:14

AI gets really, really, really powerful

20:16

and has lots of knowledge and ability

20:18

to make things manifest in the

20:20

world, then it kind of makes sense

20:22

that the God would come that

20:24

way. I haven't heard anyone personally –

20:26

I've never had anyone suggest to

20:28

me that the return of Jesus might

20:30

come in a robotic form, but I

20:32

honestly don't suppose it'll be too

20:34

much longer before someone does. In Buddhism,

20:37

you have this all -pervading presence of

20:39

Buddha, and sometimes that involves integrating

20:41

technological objects like funeral practices for

20:43

dolls or printing blocks. 40 years

20:45

ago, the roboticist in Japan, Masahiro

20:47

Mori, he suggested that a robot

20:49

could become a Buddha. It's like

20:52

a one -line almost a throwaway

20:54

in his book, just this one.

20:56

Robots have Buddha nature because everything

20:58

has Buddha nature, so someday a

21:00

robot could be a Buddha. And

21:02

in 40 years, nobody has come

21:04

out to say, absolutely not,

21:06

that's unreasonable. Having Buddha

21:08

nature is kind of what it sounds like.

21:11

It means having the seed of Buddha's wisdom

21:13

within you, a seed that can

21:15

grow with learning and practice, and that

21:17

when fully matured will clear all

21:20

illusions from your mind and lead to

21:22

enlightenment. That's kind of

21:24

the idea behind Mindar, the Buddhist robot in

21:26

Japan. It's a robotic

21:28

torso covered in silicone meant to look

21:30

and move like a real person. And

21:32

unlike Nexi, the wisdom

21:34

it offers comes from its Buddhist -influenced

21:37

algorithms. It can live

21:39

forever, and so it can

21:41

keep becoming wiser. But

21:43

what do most people really think about this?

21:46

Do they think it's great or

21:48

creepy? I want to

21:50

start with biotechnology. If you go back to

21:52

the 1970s when in vitro fertilization was

21:54

being created, Leon Cass, the bioethicist, said this

21:56

was a terrible idea. It was going

21:58

to do all these terrible things and he

22:01

knew it was a terrible idea because

22:03

when you told people we're going to

22:05

fertilize eggs functionally in a test tube

22:07

that, which isn't really how

22:09

it's done, but we were going to do

22:11

it that way. People are going, oh, that's gross

22:13

and unnatural. And so he

22:15

coined this phrase wisdom of repugnance

22:17

and the common term was the

22:19

yuck factor and he said anything

22:21

we feel yucky about, we just

22:23

shouldn't do. Now the

22:25

reality is in vitro fertilization is

22:27

now a technology that's enormously comfortable

22:29

to people, including Leon Katz and

22:31

so we get really used to

22:33

things. So I've been

22:35

asking my students about this for many,

22:38

many years now for the decades I've

22:40

been teaching and I

22:42

started asking them when I first taught a

22:44

religion and science class, let's

22:46

say you have a domestic robot that

22:48

helps wash and keep things working around

22:50

your house and it can talk with

22:52

you, it converses with you and one

22:54

day it says, may I come to church with you?

22:59

Will you let it come to

23:01

church with you? And 20

23:03

years ago I would get maybe

23:06

one out of 25 students,

23:08

two out of 25 students kind

23:10

of hesitantly going maybe and

23:12

all the others were kind of

23:14

uniformly no. And when

23:16

I ask them now, and it hasn't

23:18

been that long, but when I ask

23:20

them now as I continue to every

23:22

single year, now about half of a

23:24

class will just say, sure, why not?

23:26

And so there's been a massive change

23:28

at least among young people attending

23:30

a small Catholic liberal arts college

23:32

in the Bronx, right? So I

23:34

mean, that's my sample set for

23:36

that particular question is my students

23:38

at Manhattan College, but I've seen

23:40

really, really significant changes and I

23:42

think what has happened is that

23:44

people are used to their technologies

23:47

in a way they weren't. It's

23:50

true. Sometimes now we

23:52

light candles in church by pushing a

23:54

button that turns on an LED flame. We

23:57

have robotic pets. They're

23:59

even smart refrigerators that know when we're

24:01

running low on groceries and automatically add

24:03

them to an insta card account to

24:05

remind us to go to the store. It's

24:08

going to be all those little baby steps

24:10

right? I mean already people name robot vacuum

24:12

cleaner. That's right. And it was well known

24:14

when the when the room of vacuum cleaner

24:16

was kind of first made that people were

24:18

mailing them back broken and saying don't send

24:20

me a new one you have to fix

24:23

this one. Like this one's a part of

24:25

my family you have to fix it. You

24:27

have to fix it. It's not

24:29

clear if Rumba honored those requests,

24:31

or if like some desperate parents

24:33

who dropped a new goldfish in

24:35

the fish tank to replace a

24:38

belly-up one before their kids got

24:40

home, they just hoped no one

24:42

would notice a swap. But either

24:44

way, people wanted their robot. After

24:46

all, it usually had a name.

24:49

People will name objects, they

24:51

form relationships with objects, even

24:53

objects that are terribly hard

24:55

to form relationships too, like

24:57

a frisbee-shaped robot vacuum. So

24:59

by the time it gets

25:01

much more able to engage

25:03

with us, then it will be a

25:05

natural progression for people to start

25:07

seeing it. And I think when

25:09

I think of my students 20

25:12

years ago saying absolutely not, a

25:14

robot can't come to church with

25:16

me, they didn't have interactions with...

25:18

computers that were

25:20

sufficiently sophisticated, that they could

25:22

see that as a plausible kind

25:24

of question. Whereas now my students

25:27

are talking to their phones, they

25:29

might have something like an Alexa

25:31

spying on them at home. You

25:33

know, they've got technologies that they're

25:35

talking to and that sometimes talk

25:37

back to them, and so they

25:39

have a whole lot more comfort

25:42

and familiarity with that conceptually.

25:44

And one little bit at a time, right?

25:47

But what if robots get too good?

25:50

From an engineering and design

25:52

perspective, Cynthia says it is

25:54

possible for AI to actually

25:56

be smarter than humans. But

25:58

what does that mean for... religion

26:00

and really for all of humanity. I

26:03

think this is kind of this

26:05

provocative story about AI right now.

26:07

Let's call it narrow AI, right?

26:09

So once you kind of understand

26:11

what the thing is to optimize,

26:13

you can run these algorithms to

26:16

do that potentially in a way

26:18

that far exceeds what human capacity

26:20

is and to outper people, right?

26:22

I'm wondering what you think about

26:24

that because in the right hands.

26:26

Those can be amazing tools, right?

26:28

I know you design robots to

26:30

keep kids comfortable in the hospital

26:32

and make them feel at ease,

26:35

but you can build this sense

26:37

of trust and empathy and camaraderie

26:39

with robot by doing these things.

26:41

But if you're a marketer or

26:43

you have some other intent of

26:46

trying to convince people of

26:48

something, doesn't that also give

26:50

you tremendous power? Here

26:55

I am at the Media Lab building

26:58

these robots that through social interaction as

27:00

you're saying, yes, I mean, we're talking

27:02

about helping someone learn or helping someone

27:04

adhere to a diet and exercise program

27:06

or to be more emotionally resilient. I

27:09

mean, these are AI systems that are

27:11

shaping our thoughts, our beliefs, our behaviors,

27:13

I mean, all of these things. Now

27:15

we're doing it to try to empower

27:17

people to make better decisions for

27:19

themselves, but there's another whole other side

27:22

to that coin, right where it could

27:24

be used to manipulate. Systems are

27:26

capable of persuading you in potentially

27:28

really profound ways. And you have

27:30

to ask the question, so who's

27:33

persuading me for what reason? And

27:35

of course as human beings, we

27:37

try to persuade each other all

27:39

the time. It's just part of

27:42

social interaction. But now you're

27:44

talking about an entity where

27:46

there's layers of that, who do you

27:48

trust? That's a good question. When

27:50

it comes to asking Syria or Alexa what's

27:53

on the shopping list, or what the

27:55

weather's going to be tomorrow, the stakes

27:57

aren't that high. But when it comes to

27:59

religion... Questions about

28:01

morality, about meaning. The stakes

28:04

can get pretty big, pretty fast.

28:06

The thought of a robot becoming

28:08

a charismatic leader, like a

28:11

politician or a Pope, is kind of

28:13

chilling. I think at this present

28:15

stage here in the early 21st

28:17

century, there are lots of people

28:19

watching these technologies develop and feeling

28:21

fear. Right, but we human beings,

28:23

part of the existential state of

28:25

humanity is the terror of history,

28:28

right? Like, we're always scared of

28:30

the things we did in the

28:32

past and the things we might

28:34

do in the future. Those warnings

28:36

we get from pop culture that

28:38

say, we're going the wrong way,

28:40

our technologies are going to... overmaster

28:42

us and destroy us, that's really

28:44

a fear about us. That's not

28:47

really a fear about our machines.

28:49

That's the recognition that we're not

28:51

always doing right in the world

28:53

and that our obligation is to

28:55

do right. That's a critical point.

28:57

Yes, someday soon, AIs might have

28:59

an almost god-like intelligence, and

29:02

given the number of cameras

29:04

we have around, a god-like

29:06

omnipotence. But it's we who are

29:08

creating those AIs. It's we who, in

29:10

a sense, are their gods, at least

29:12

right now. And so if we want

29:15

our creations to be good,

29:17

to be virtuous, we have to teach

29:19

them how. We have to imbue them

29:21

with a moral sense. If we

29:24

don't, if we rely on letting

29:26

them learn just by conversing

29:28

with people on the web,

29:30

they're going to end up

29:33

like Microsoft's infamous AI chat

29:35

batte. who within days turned

29:37

into a misogynistic racist. And

29:39

so, in a sense, robots need religion

29:41

or rules of ethics as much as

29:44

the rest of us, which is why

29:46

we scientists and designers and academics

29:48

are all putting our heads

29:50

together. There's room for philosophical

29:52

ethics, there's room for biological

29:55

ethicists, for religious ethicists to

29:57

come together and talk about...

30:00

What ought we be doing around us

30:02

and how do we fix the problems?

30:04

And then if the folks in

30:06

the industry say, yeah, we want

30:08

to draw on that. We want

30:11

to build machines that are

30:13

actually actively helping, then I think

30:15

we'll get there. But as with any

30:17

religion, you have to be

30:19

careful. We have to be clear and

30:22

honest that religions are not

30:24

always good for people. Sometimes

30:26

people do terrible things. in the

30:28

name of their religions or even

30:31

provoked by their religions. But people

30:33

also do beautiful, wonderful, amazing things

30:35

in the name of their religions

30:38

and provoked by their religions. And

30:40

so an honest appraisal might say,

30:42

okay, if I'm looking at, say,

30:45

in Hinduism, the darma, or which

30:47

is usually translated as duty, if I'm

30:49

looking at duty in Hinduism and thinking,

30:51

what does that mean for... human to

30:53

human and human to machine relationships and

30:56

how do I want to design AI?

30:58

I don't want to take Darma and

31:00

go, okay, that means that some people

31:02

are born with particular duties that they're

31:05

stuck with forever. That's a terrible use

31:07

of it, right? But I do want

31:09

to think about what are our sense

31:11

of mutual obligation. So what are my

31:14

duties that I owe? to others, my duties

31:16

to the environment, and we draw

31:18

on our religious traditions and practices

31:20

to do that, that would be

31:22

a wonderful thing. And so I

31:24

do think there's a lot of

31:26

space for conversation in

31:28

the AI ethics field that it's not

31:30

just about... philosophical positions on ethics, and

31:33

of course there's a place for that

31:35

too, but also religious ethics and religious

31:37

actions. What is a good Christian, a

31:39

good Jew, a good Hindu? What do

31:42

they ought to be doing in the

31:44

world? And once we look at what

31:46

they ought to be doing in the

31:48

world, we're going to start looking at

31:50

kind of overlaps that produce probably a

31:53

good ethos for what we want our

31:55

machines to do in the world also. So

31:57

in some ways, it's not just us

31:59

learning to worship in AI

32:01

or a robot, it

32:03

is spiritually developing that AI itself,

32:05

giving it that spiritual in

32:07

some sense in terms of principles

32:10

grounding so that it is

32:12

as it grows and develops, it's

32:14

going to have the right

32:16

kind of tealess or purpose of what

32:18

it should do. Yeah, no matter how

32:20

complex it ever gets, whether it gets

32:22

to human equivalents or greater, if we're

32:25

building that in, if we're leveraging those

32:27

religious values to benefit the most marginalized

32:29

to create a better world around us,

32:31

if we can do that, we will

32:33

have served the future well, regardless

32:35

of how good the robots get. So Robert, as you

32:37

think about maybe 20 years, 30 years down the

32:39

line from where we are now, and I know it's

32:41

all speculation and we're not sure how it's going

32:43

to come out, but what ways

32:46

do you see maybe

32:48

technology, AI, robotics influencing the

32:50

average person's worship? So

32:53

my hope is that people would find

32:55

that the technology, that you get these

32:57

initial hype waves right and everybody jumps on

32:59

the technology, and then they come to

33:01

kind of realize maybe I don't want to

33:04

be on Facebook all the time or

33:06

whatever. And so we get these up and

33:08

down kind of waves of use and

33:10

it may be that there are

33:12

better and worse times for

33:14

people to engage in something

33:16

like a robot at home

33:18

that could get quite interactive. I

33:20

mean, you could imagine it getting

33:22

quite sociable in that fashion,

33:24

but you could also imagine it suggesting

33:26

you go to church with other

33:28

people once in a while. And

33:30

that might be a good kind

33:32

of programming element in all of

33:34

this might be if it's coming

33:36

from a spirit of genuine religiosity,

33:38

maybe the robots, the apps, whatever

33:41

could say, hey, you're doing great.

33:45

At heart, much of religion is about

33:48

connection, a caring

33:50

connection with the divine, but also

33:52

with our fellow humans. And

33:54

it's here that AI driven

33:56

robots could be a big

33:58

help. They can be a

34:00

spiritual companion or even God.

34:02

for people who can't go to church or temple because

34:04

of being homebound, or in areas where there's a shortage

34:06

of clergy. Or like an advanced mind

34:09

are, they can offer inspiration to

34:11

whole communities of worshippers that

34:13

come together to see them. In some

34:16

ways, they might even prove to be

34:18

superior preachers, less vulnerable to motives

34:20

for power or bias than human

34:22

religious leaders can be. In the

34:25

best world, they'll help us

34:27

grow, spiritually and otherwise. If,

34:29

and here's the big if, we design

34:31

them to be good partners. To

34:33

me, that's really the question of,

34:35

it's the partnership. That's what I

34:38

care about, and how do you need

34:40

to design these entities that are engaging

34:42

and natural for people to interact

34:45

with so that they can empower

34:47

us and complement us? What do we

34:49

as people need that's going to

34:51

help us be who we aspire to be

34:53

to live in a society we want to

34:55

live in? How

35:01

God Works is hosted by me, Dave

35:04

Disteno. This episode was written by

35:06

Josie Holtzman and me. Our senior producer

35:08

is Josie Holtzman. Our producer

35:10

is Sophie Eisenberg. Our associate

35:12

producer is Emmanuel Desarmé. Executive

35:15

producer is Genevieve's sponsor. Merritt

35:17

Jacob is our mix engineer

35:19

and composed our theme, which

35:21

was arranged by Chloe Disteno.

35:23

The executive producer of PRX

35:26

Productions is Jocelyn Gonzales. This

35:28

podcast was also made possible

35:30

with support from the John

35:32

Templeton Foundation. To learn more

35:34

about the show and access

35:37

episode transcripts, you can find our

35:39

website at how God Works, all one

35:41

word, dot org. And for news and

35:43

peaks at what's coming, feel free to

35:45

follow us on Instagram at How God

35:47

Works pod, or me on X or

35:50

Blue Sky at David Disteno.