at Stanford University, and a Race and Technology Affiliate at the Center for Comparative Studies in Race and Ethnicity. Her work is informed by the understanding that racial perception and identification impact outcomes for health, wealth, and social justice. Her most recent book project, 'Race Making in the Age of AI,' considers how the humanities and arts function as crucibles through which to frame and address urgent social questions about equity and social justice in socially transformative technologies. She teaches what sounds like fascinating interdisciplinary courses at Stanford, ones I really wish I'd had the opportunity to take when I was an undergraduate student. And I get to work with her closely at the Institute for Human Centered Artificial Intelligence at Stanford. Michele, I'm so glad to have you on the podcast today.
I'm so happy to be here. I look forward to talking with you.
and how you got interested in the things that ultimately led you to choose the career path that you are on, or maybe what you didn't choose and what just sort of organically happened. So, tell us a little bit about your childhood and how you got onto the path that you're on.
Thank you. I feel like we're already in a therapy session, because the path chose me. It's so odd, really, how I got interested in this. I think about this all the time when I'm teaching my students who, it always looks like, they have a straightforward path to a major and then a career. But mine was really peripatetic. So, I was a humanities person. I wasn't a technologist. I wasn't particularly interested in math. I wasn't bad at engineering.
because I remember him. He was a straight up engineer, and he hated that the government was sending monkeys into space. He thought that was completely unnecessary. He hated the politics of the Cold War --
the sort of politics informing. And at that time, you remember, very early AI, I think he was -- it was in the 60s and 70s, it was mostly government contracts. He was working for Control Data and then very early artificial intelligence. And I was completely the opposite. I don't know whether it was just having a father for whom the world was a problem to solve. There was a whole engineering mindset that had full explanatory force for everything.
but right before I joined the Institute for Human Centered AI, or thereabout. And he would have been flabbergasted that I was sitting around the table, as you know, that that institute, at least at the time I joined, we had eight Associate Directors of HAI, it was very interdisciplinary intentionally. So, Fei-Fei Li and her -- John Etchemendy decided that they wanted, in addition to computer scientists and engineering, I was representing arts and diversity, humanistic perspectives. We had someone in the medical school, someone from law. She'd taken sort of C.P. Snow's, 'Two Cultures' and blown it up and she really wanted people sitting at the table.
always seemed unintentionally, I'm sure by him or culture, to set technology and technologists above the arts, which were always seen as sort of ornamental, nice. Maybe they add local color to a product or something. And so, I was always frustrated by the different cultural statuses with that. So, I'll just say it is bizarre that I'm sitting here talking to you, Kevin. And I think we, if I'm not mistaken, we actually met in Rome.
The dialogues and that was really what started I would say quite late in life obviously, my interest in this.
Father Eric Salobir and other people, who were talking about spirituality and AI and not in an apocalyptic or hallelujah way, but what are the bigger implications? And there was a conversation about storytelling, if I recall. And the people in the room, I think were thinking storytelling like marketing. How can we tell this story better? And I was like, 'There are so many more narratives about technology than 50s science fiction by white cis-male authors. I mean, that's there, too.'
And I think we connected over that too, because you have such an expansive imagination. And then I think the other strange thing that happened,
and unbeknownst to me, had sold my image to Getty Images. The university didn't - I wasn't at Stanford at the time - didn't know either. And I started seeing hundreds of pictures of me out there. But I started being lightened and darkened.
And my father-in-law saw me on the side of a Boston bus as the older than average non-English speaking student going back to school. I was on Harvard's Rainbow site. I was on the Hispanic Congressional Caucus website as a proud Latina moving in. So, I became very interested in one, the appropriation and commercialization of people's images. And I remember you telling me at the time, you're like, 'That is nothing.'
But both of those set me on a more cynical path, I think, to thinking about AI and both because I was cautious because of my father's true believer commitment that engineers would solve the problems of the world. And then also, the use of my image and that weird sense of losing control of your identity.
Well, look, and I think it's perfectly reasonable to be cynical. I mean, it's sort of funny, like it's --
Yes. Yes, cautious, right, or skeptical or--. And I think as an engineer you kind of have to be and it sounds like your dad was very much this way. So, like you, yes. By virtue of your acculturation or the brain that you were born with, you kind of look at the world in terms of everything that's broken and what you can do to go fix it, right? That's the impulse of engineering. And it's like a pretty jaundiced way to look at the world.
I like the way you described it, though, because that to me seems also like an ethical way to look at the world. Because I remember my -- it can be limited if we don't pair it with other ways of looking at the world.
Through different vocabularies and mindsets, yes. But the way you just described engineering actually sounds very
which I realize is sort of antithetical to general intelligence, but he was very concerned with the social impacts of the things that we create and the social implications of them. And that's what I hear when you say that then, too. There are problems in the world that engineering can address. So, I definitely appreciate that.
I want to go back to your father in just a second and talk about why it is that you think he may have held engineering and technology apart and above from the humanities or other things. And you and I both know in academia that there's always --
about being a human is like we're all just playing our role. Hopefully we've chosen to do the thing that we're especially good at. And hopefully we've all got enough wisdom to understand that we don't know everything and that we need all of these other perspectives to come in and help us, particularly as engineers. You can sort of look around and see that everything is broken, but maybe you don't understand why. Maybe you don't understand that not all fixes to all problems are equal in merit. Maybe you don't even understand what's worth fixing and what isn't.
perspective on things is just mindset that you're born with, because I do think that there's some really particular things about an engineer brain? And how much of it is it sounds like he's a Cold War, post-World War II trained.
And so, a lot of it culturally is like, 'Hey, we're in this existential battle, and science and technology is the way we win. And here's the stuff that's important.' And yes, it's easy to be in the middle of that and you think that you're more important than you actually are.
He didn't quite have the hubris. I think it was -- it was hopefully --
if I can give this example. Well, this is -- it was a sense of security and control. I was listening to --
a couple of your other interviews with Mira, for instance, Murati, and the sense of -- especially when the world is in chaos, it seems like a way of making sense of things that bypasses politics and bypasses sort of social issues, all the things that at least I was growing was particularly interested. I remember to his great credit, he always wanted, I was the eldest, me to not take sewing class, which is -- we had -- I can't remember what it was called. Like Home Arts or something. And then -- or you could figure out how to fix a car engine.
So, they're just coming into their own and their bodies and he was like, 'Here's the hammer and here's the nail and the -- and here's a screw and a nut.' I can't remember how he described. And he's like, 'One is the male part, and one is the female part.' And I was just like, 'This gendering of these tools is not -- is a problem dad.' And I was feeling it like kind of horrified and -- that he was talking about male parts and female parts and female parts were passive and male parts were active. And even at 12, 13, I could see the gendered implication. For him, it's just like that's just how it is. And so, it didn't take him to thinking about the way his own field incorporated social values. It was meant to be objective and neutral. And so --
to just learn all the technical aspects and then belatedly maybe after you went into industry, you got to ask questions about the ethics or application. Students want to know now. They want -- they're interested in social psychology. I was just having a student talk to me about the Weizenbaum's Eliza Effect in relationship to the arts, and they want to integrate those questions about social psychology, whether or not they go and build a product, but early on. And I think of my dad's era. It was very much, 'This is the domain of another field. You can deal with the social issues. We'll deal with the technical one.' And I think he associated the social with political that he didn't want to -- that he also found muddying pure math and muddying pure --.
and to your point, we need multiple perspectives. When you sit down at the table, which literal table, and I was luckily at HAI, we call it, as you know, the Institute for Human Centered AI. We were all equal partners. However, the vocabularies, the way we talk about the world, the things we prioritize, the way we document our work in a universe where the quantitative is so important, I found, I know my colleagues did too, the way to legitimize humanities and arts was often by -- through quantitative methods. We would start to use the same language and metrics that the engineers were to measure success or impact and things like that.
Like it just -- the arts move us and they're -- we had very vague language or at least the way we talked about its potency, the arts potency, didn't compute literally with some of my colleagues. So, I have come to really appreciate how hard it really is. And there's outsized money in technology in ways not in the arts. What it means to come with goodwill together and try and hear each other when we have studied differently, learn differently and sort of speak differently --
about the things that are important. I always find it -- my father used to say and then we'll get off my father.
The artists on the one hand and art was often seen as sort of ornamental.
as sort of incidental, and he bemoaned that. He thought it should be more important. But then he also pointed out that the first thing that fascist or authoritarian governments do was get rid of the artists, because they're the ones that shape the cultural imagination. They're the ones that actually -- and that's very powerful. At least he saw they're both seen as irrelevant, and yet a primary threat. And that always fascinated me.
I mean, it's always, I mean, like maybe this is getting a little bit too philosophical and squishy.
You're creative. But you also have things that you do that you're an engineer. You're a problem solver. You have to use quantitative methods for things. And so, it's just kind of -- what context are you in and what skill do you need to be able to manifest to solve the problem that you're working on. And like we get a little bit --
But do you feel like --? Well, I'm super -- you're total -- I mean, I think I know where you're going, but I would love that in that world, too. You authorized yourself to create. You do sculpture. You do clay, right? You work with clay?
I only recently found this out about you because your Instagram post is amazing. But there's a case where you took yourself seriously enough to do that. And I imagine if you had decided to go into that as a --
So, yes, there's one thing that -- there's this joy that I find. I just have curiosity about everything and it's thrilling to go wallow in that curiosity. But there, especially I think when you're young, there's a risk to casting the net too wide because there's some things that you want to do that are complicated and you have to focus for a little while to get good enough at a thing so that you can bring something to it. It was hard to learn how to program well. It's hard to do a bunch of things and you just have
This is going to be my professional identity. I have to focus enough to get into the, I don't even want to call it the elite, but I just want to be able to at a high level, go participate with a bunch of other people who are working at a high level on this. And I want to be able to understand what they're doing and contribute. And yes, I'm not even saying that that's the wrong thing to do. It's sort of the advice I give to my kiddo who my 16-year-old is a bio nerd, and she also is interested in a gazillion different things. And I'm like, 'Look, you have to -- at some point, if you care enough about this one thing, you have to invest enough of your time and energy
like hundreds, that tension between having deep expertise and taking the time, the duration to get good at something, whether it's in the arts, whether it's in literature, whether it's in engineering is challenging because for instance, I've always taught at liberal arts institutions. And that -- it means that you have a level of expertise, but you are simultaneously also learning these other things.
I don't know if you recall, but a few years back, the engine -- the credentialing for engineering, our engineering majors, is external and if they could, they would grab these kids in utero and start training them. I mean there's no profession including my own that wouldn't say if you just laser focus on this
real expertise to excel, not just get by, and yet also, bring in with enough and not just casually like an elective, an arts practice elective, and somehow students manage to do it. We've been thinking -- some of that's a curricular thing. Can we make it so that we hold at bay whatever field of expertise it is that's saying, 'No, I'm going to train you for this profession,' which is problematic, not least of which because it will change. But also, you said something earlier that I thought was so important. So, you obviously had internalized this earlier, which was you wanted to create things that had it -- that changed the world.
That we're makers from the beginning, but I actually think we're storytelling sapiens because we are --
but they're not so far apart, if you think about it. Not to flatten makers and we're all makers or creators, which is to -- in some ways do a disservice to the expertise you're talking about. Also, that is essential. But I feel like we don't -- to go back to your thing about we don't have to have categories. That's an example where we don't have to simply prioritize. We don't have to typologize -- create typologies of what we can do. I think that -- talking with you is always fascinating. I like the squishy philosophical, actually, because it has real practical implications when you're in school, just as you said, because a lot
if school doesn't kill the joy of curiosity, want to know things, want to learn things. But you mentioned and I was thinking about this a lot when I found out that you do pottery. If you call it pottery or do you call it pottery, the work that you're doing --
which is beautiful. There's pictures of the kilns and the thing about arts that I would love brought into technology more, and the woman I taught with before was a MacArthur award-winning artist, Camille Utterback, who reminded me of this too, is that the practice of art making is durational and iterative and recursive, and it takes time in that kiln. And so, what I have partly --
right? You're saying you're agreeing. So, I'm lucky you're agreeing, but I think it seeps into the way we talk about things.
No, I violently agree, and I think what the world needs from us is largely contextual. So, I was in Japan a few weeks ago and no surprise, I spent -- a very interesting trip. So, I spent the first half of the trip opening up Microsoft Research Tokyo. So, a new research lab --
and entirely focused on AI. And then I spent the last half of the week talking to a bunch of Japanese craftspeople, so potters,
and folks who make canvas bags. And so, the vast majority of the people that I talked to on the last half of that trip have been working inside of the tradition that is their craft, their entire lives. And like the tradition itself is centuries old. And it's super interesting. Like --
How do you reconcile those two things?
Well, I think Japan does, as a society, does a very good job reconciling the two because on the one hand they are more progressive than most places in the world, including in some cases the United States and adopting very, very advanced technologies.
they make no sense whatsoever.
But in fact, they make a lot of sense inside of Japanese society because if anything, they sort of look at the technology as purely a tool, an instrument that's in service of Japanese people. And the art is something that's in service of your soul or your spirit or --.
Right. Right. Yes. I love that. And I love that you said the word 'revere.' And the other thing in that description that makes me realize I need to go to Japan is that so much of AI is so future-oriented. It's always -- it's like unprecedented supposedly, and never been seen and it's always --
past and history as if it was more remedial or something we have to sort of get over. The way you just talked about Japanese culture as probably even revering the arts in that sense, even more, and tradition is beautiful. And you mentioned our AI is a tool, too. One of the things that has come up a lot is the idea that it's a tool, but it's not exactly neutral. I don't know how you think about that, but it does embed social values the way language does, too. So, I love it as a -- in its proper place as a tool in service of it. I think sometimes we maybe say 'AI for good'
or any other tool neutral is almost like a little bit of sophistry. It's kind of dancing around the issue that you really are getting at because it doesn't matter whether the piece of technology is neutral or not. Ultimately, a human being is going to go decide what to do with it and human beings are --
Yes. Well, I mean it's interesting because the implications of that do matter. So, either it's in -- when I mean -- so, if you think that it's bad actors, that it's not the technology, it's the misuse of technologies, that's one approach to solving technologies you might not like to see in the world.
Who's in place to think this is valuable or not? And so, those you know...
And it's one of those things too that's context dependent. One of the first projects that I worked on when I left academia and joined industry a million years ago was I built very early machine learning systems that made judgments about which content was adult or not.
Yes. And so, you will find very quickly that that is very context specific. So, what one country or society or culture finds offensive --
another may find acceptable.
But it's sort of interesting and yes, it's eye-opening when you have to sort of confront these things like as an engineer because I was 30 years old at the time. I was a little bit like your dad. It's like, 'All right, well, there's a reductive way of looking at all of this stuff and the algorithm is going to be simple and it's going to -- yes.
But I love that example. I think that's so moving because beyond just content moderation, but at HAI, they're doing work right now about if you recognize that certain technologies that were created amplify whatever bias, this is so not new to you but -- or social agitation or whatever, then --
who are the moderators of it. But you can see within a heartbeat, if you're trying to do good, it puts you right in the realm of social psychology and ethics and I think the best engineers and you are so one of them, are like, 'Gee, maybe I need to know more?' I think that's important. I think -- I won't name the names, but there's been a few books that have come out and said, 'Well, humanists, people or laypeople just need to know how the technology works more. They really -- they're not credentialed to weigh in on policy or anything else.'
Well, given that it impacts every aspect of our public and private life. But I felt that way, too. I felt like at HAI before I -- I didn't want to just start critiquing something that I didn't technologically understand.
there's like huge debates and fascinating research out there, not critique meaning critical, but that can enlarge and encourage, I think, engineers or whoever happens to be creating, the arbiters of the worlds in which we live to think about the things the way you are, the way you always have, Kevin. You've always been like this, as far as I can tell.
I wish one of the things that we taught our young engineers was for them to lean into the sort of discomfort that arises very quickly when you're trying to build technology for human beings. And they lean into all sorts of discomfort, right, because you walk into your first partial differential equations course and like --
