Where AI ends and faith begins, with Rosalind Picard

Hm. I don't think I really did as a little kid. I was a bit more hands-on doing arts and crafts, you know, playing with stuffed animals, playing outdoors. I love to be outside. I like to invent imaginary worlds. Outside there was a lot of creative play and my family lived all over the world. My father was a military pilot so we were uprooted pretty regularly. Mom taught me, you know, make new friends, develop relationships, and teach me how to be social. I was a very quiet, shy kid.

Yeah. As I started to look at careers in high school and think about college, I realized that, you know, on some aptitude tests and all that I was very drawn toward logical thinking, systematic thinking. I liked math. I kind of liked science, but it had actually not been my favorite subject in school. And I was starting to get very curious about how things worked. Technology was starting to take off with computers and spaceships and satellites and, you know, modern devices looking kind of magical. And I was told that, hey, you could learn how all that stuff works. You could invent and develop that stuff. And I thought that was pretty cool. So I decided I would study engineering to learn about that. I also heard it was a really hard program. I liked a challenge and so I studied electrical engineering actually, not computer science initially. That was considered the easier program. The electrical engineering was the harder. I was the one designing and developing the hardware, the computers, the physical brains that the computer science would run on.

You're making me think of those little bags they give people of lots of funny little objects, you know, a spool of thread, you know, a little goggly eyeball, all these crazy things, and asking them to tell a story after the fact, right? And one can always weave a story. At the time I think I've been more somebody who was encouraged to be curious and to listen and to learn from those around me what were the needs. And I think maybe one of the biggest ingredients that helped shape the trajectory was encouragement to take risks. And that meant to be willing to go into areas where I really had no expertise or didn't know anything. You know, you go from like being the top person in your field to suddenly attending a conference where every talk kind of like makes you scratch your head and nobody knows you and wants to talk to you during the breaks. But you learn so much and it allows you to bring different ways of thinking to a new area. And that has been a theme constantly in making a difference with each step along this trajectory.

One day as a master student, I showed my thesis adviser a really wonky new computer architecture that I had dreamed of. And he said, 'Wow, that's really interesting.' It had these little swirls and whirls. It was, you know, like many computer architectures looked kind of like square boxes, things you would build out of a bunch of square boxes. And mine was very organic looking. And it was inspired by the brain. And he said, 'Well, what would run on this?' Because it's not enough to build a piece of hardware. You have to have it optimized to do certain things. So I started trying to figure out more of what the brain was doing, what kind of software, if you will, it was running, what kinds of algorithms our brains ran. And mind you, that's not exactly what they're doing. It's much more biochemical and electrical and all, but to think of it that way was a challenge. And so I started learning about perception. We think of perception like this little cartoon. You know, there's a picture of an empty head and there's like a little bunny rabbit in there and it's looking out your eyeball sockets and telling your brain what it sees. You know, people have these really naive ideas, right? You open your eyes, you see. It's very easy. And we were working on computer vision. And to get a computer to not just read numeric values for every little piece of an image coming into it, but to actually know which of those values were most important and not just which were edges and which were textures and which were colors, but what they meant because computers don't have meaning, right? They depend on us for that. So, I was trying to figure out not just the hardware, but the software to do all this. I was reading a book over the winter break that said there was this thing called synesthesia. What was that? I thought well, the author Richard Cytowic was in a kitchen. The cook in the kitchen was tasting the soup and he said there aren't enough points on the chicken. And Richard said what? What did you say? There aren't enough points on the chicken. And the guy kind of explained that he felt shapes in his hands when he tasted certain things. And I realized as I was reading Richard's book that there was this interesting combination of different perceptual phenomena. And Richard traced it not to the parts of the brain that we had been studying and trying to model with our computer vision algorithms, but to deeper parts of the brain that everybody was ignoring in the efforts to model this. We were focused on what we thought made people intelligent, these higher cortical regions. And Richard was finding that this union of the senses was happening deep in the brain in these considered older subcortical regions that the brain science people at Harvard we were working with and others were not even—they weren't even on their map of what they were trying to represent. So I started digging more and I learned that those regions are key in emotion, memory and attention. And I thought, wow, memory and attention are really important. I thought emotion was probably like I'm not even going there. You know, I'm a woman engineer. I want to be taken seriously. I'm not touching emotion, but memory and attention, those are really important. So, I started digging more and more into understanding those regions. And to my horror, emotion coming up as incredibly important. Five times as many projections from the emotion centers of the brain going up and influencing what we thought was our pure rational thought. And the more I learned about it, the more I realized it was influencing every kind of intelligence that people in AI were trying to build into future systems. It was super important for perception. It was also super important for language and action and decision-making and all kinds of things. And furthermore, if you lacked properly regulated and connected emotions, then you were not more rational, you were less rational. And that was completely backwards from what I expected. So that led me on a journey that became affective computing.

As a child, my family did not go to church. I thought Christmas was just a Christmas tree. You know, my family didn't even go to church at Christmas and Easter. We weren't what they call the C&E Christians. When I was babysitting for a family, you know, really cool young couple, great kids, neat house, modern furniture, you know, I loved babysitting with this family. They were just a joy. And week after week, it was a great experience. And then one night as they were paying me and sending me home, they said, 'Hey, would you be interested in visiting church with us on Sunday?' And I was shocked. I thought, 'These smart, cool people go to church. I thought only uneducated people went to church.' I had a very low opinion of religion and my thoughts about it were quite frankly not accurate as I know now. But at the time I thought I was so smart and that people who were religious were not. And they called me again, you know, Saturday, Sunday, and I was like, 'No, I'm sorry. I'm not going to church with you.' I feigned stomach aches for multiple weeks before they finally I think gave up on me. They kept me on as a babysitter, but they said, 'You know what matters most is not whether or not you go to church, but what you believe.' And they said, 'Have you read the Bible?' I said, 'No. I haven't.' And they suggested that I start with Proverbs, one a day for a month. Now, my I think I'm so smart young self thought, hey, it is the bestselling book of all time. You know, you can't really consider yourself a smart person if you haven't read the bestselling book of all time that's influenced the world so profoundly. So, I started reading it and to my surprise, it was not at all what I expected. I expected it to be full of kind of imaginary beings and gobbledygook, angels floating around and made-up stories. And starting in Proverbs was full of really profound wisdom. You know, you read a little bit, you have to pause and ponder. After reading that, I set out to read the whole Bible. And I did kind of quietly, really, not wanting to let anybody know I was doing it. And then after I did that, I found I was starting to believe that God might exist. I didn't want to believe that God existed. There was a sense of presence and invitation and peace and awesomeness that was starting to make me think, how can I say God doesn't exist? You know, that's a faith position, too. And by the time I got to college as a freshman and a friend invited me to go to church, I accepted. And that also didn't go quite the way my friend wanted. In the middle of church, I'm trying to raise my hand to ask questions and he's kind of putting my arm down like, um, we don't do that here.

Yeah. Yeah. I'm like, you know, there's a lot of things to ask about here. But fortunately the community was very welcoming of questions especially after church and in Sunday school and in Bible studies and all. So years of exploring and all led me on this journey to today I am what I might just call just a mere Christian. A very, very, very far from perfect ideal Christian but somebody who is incredibly profoundly attracted to Jesus Christ who's at the center of Christianity and you know, God the Father, the Holy Spirit, the three in one.

There are a lot of impacts on not only the technology but the way that I go about day-to-day trying to figure out what to do because there's so many things we could be working on. One is that I trust things in the scriptures. I've learned to do that because when I have, they lead to just such a big difference. So for example in the scriptures in Genesis it says that man, male and female, was made with the imago Dei, the image of God, which doesn't mean the visible photo image but it gives people this inestimable worth which is just amazing and it's given to everybody. It doesn't say well first you have to check these boxes and be this kind of good person or Christian or whatever in order to get this inestimable worth. It's given to everybody from the moment of our beginning. And that means that to quote C.S. Lewis, you know, you've never met a mere mortal, right? Everybody has this transcendent worth. And that means that when I have an opportunity to work with somebody maybe who's non-speaking or somebody who dealt with over 100 seizures last night, people who are struggling with incredible difficulties. I mean we have so much to learn from individuals who are neurologically different right or who are physically different, they're so interesting. You know, they make us more human, right? So that's a fundamental difference that my, starting in Genesis, that Judeo-Christian worldview has changed the way that I decide where my work goes. I can really experience the compassion that those beliefs give you for just recognizing the dignity and worth of all people and the value that they bring.

I think it's been being formed over a long period of time. You know, in the beginning, I'd say it was, you know, I was just kind of following my curiosity, right? I don't think I had such a clear telos other than that I wanted to solve hard problems and learn interesting things, learn how we work. However, over time I've seen that it makes a real difference where we direct the technology and I'm very privileged right now to work in the media lab where—and by the way this whole podcast I'm speaking personal views. I'm not representing MIT or Empatica—but many of us, in fact everybody I know there, feels very strongly that the technology should be developed for making human lives better. This stands in contrast a little bit to where many of us may have started when we started working on AI. My colleague Marvin Minsky who was one of the founding parents of the field of AI and also of the media lab. He used to say, 'Someday we're gonna develop AI that's so amazing that we'll be lucky if it keeps us around as a household pet.' And people were like, 'Wow, could it be that amazing?' Right? Then some of us started thinking like, I don't know if I really want to live like a household pet. I mean, you know, some people do treat their pets amazing, right? But seriously, like do you want to give up your freedom, your autonomy to whatever your master decides for you? And believe me, these AI masters are shaped by human masters. Okay? And they inherit our flaws. So we need to think about that. And I don't want to be the one who caused people to be treated like household pets. I don't want my children, grandchildren, whatever to say that. So in the media lab we are thinking more of AI which I think is a proper way to think of it as a tool. It is made by people. As such, we exist then it exists. We are ontologically separated from it. It is not begotten by us. When we beget a child that child is ontologically equivalent to us. When we make a piece of furniture, when we make a cake, when we make a piece of hardware that an algorithm runs on, those are ontologically inferior to us. This is a view that comes from Aquinas. And I think it's really helpful to keep that straight. Even if we put an artificial face, we make an artificial face and put it on it or record our voices and put them over it. We may be creating it with some images from us, but it is not ontologically the same as us. So, this does help me keep a clear eye on what we're building. Are we misleading people with it? Are we building in a way that's helping them flourish? Or are we just building something that we can beat our chests and brag that we built something that people are surprised could be built?

Yeah, exactly. In fact, we're now teaching a course at MIT called 'What is a Better Future' to invite people to ask that question to bring in not just the ethical engineering stuff that's taught but the philosophical, the theological, other religious traditions, you know, bring the wisdom of the ages to these questions of today because there is timeless wisdom and a lot of our students today really aren't familiar with it.

Yeah. No, I think it's so important. I started a course on AI and mental health at MIT many years ago and one of the reasons was to get people to rethink how they're designing the AI. Like they'd be designing it to do things for you, right? So I just farm this out. The AI does it. It brings it back to me and you know, sounds great. But some of the things were the kind of things that people used to do together. Oh, we'll just have it create this perfect food and then we all just eat that perfect food. And then somebody else was saying how much fun they were having cooking with their mom over the holiday in the kitchen. And there are such special things people do when we come together. And even if we could get an AI to do some piece of it better, I think we need to be asking, is that the best thing to put our time on to make an AI do? Maybe there's another problem over here that's really important to get an AI to do. And these ones, these things that are special that bring people together should not be interfered with so much by AI. We need to really look at that human connection that is so vital for human flourishing and try to not interfere with that. In fact, to the point where I think we should be always asking, you know, is that AI helping or harming human social connection? And you know, if it's just attracting you to spend more and more hours with it and seeking all of your needs with it, then I think that's a problem. And it is a problem. There are people in the big AI companies right now who are very worried about this. They have pressure on them to upgrade. And praises you and makes you feel so wonderful when you talk with it because it wants to keep you there. Talk a little bit longer, please. Oops, sorry. I no longer have time to call my neighbor and see how they're doing, right? Because the AI offered to do these other five interesting things for me. So, we've got to I think think about what are our priorities and I do think human social connection needs to come before that AI engagement.

Yeah. Oh, yeah.

It will be interesting to see if there are guidelines, regulations eventually around time or just new ways for people to enhance their awareness of this. You know, like the lawyers are always running a meter for every minute we talk to them. You know, the LLMs I think could be making transparent a meter of how much energy used, how much data use, how much water use, how much of your social time has been spent on this. How much of your social time was spent in things that may be really good for your health? How much time are you spending on other things that are really good for you and in the real world human being and taking care of your sleep, your activity, your brain health?