Science Friday - Imagining The Future Of AI / Face Mites. Nov 29, 2019, Part 1

Episode Date: November 29, 2019

What can science fiction and social science  contribute to how we think about our algorithmic present and future? Science fiction writers and Hugo-winning podcast hosts Annalee Newitz (author of The Future Of Another Timeline) and Charlie Jane Anders (author of The City In The Middle Of The Night) talk about their work imagining future worlds and new kinds of technology—plus how all of this fiction traces back to the present. Then, AI ethicist Rumman Chowdhury joins to discuss how social science can help the tech industry slow down and think more responsibly about the future they’re helping to build.  Plus, everyone has face mites—including you. But they have a fascinating evolutionary story to tell. In this interview recorded live at the Sydney Goldstein Theater in San Francisco, Ira talks with entomologist Michelle Trautwein of the California Academy of Sciences about why face mites live in our skin, where we get them (spoiler: thank your parents!), and how mite lineages can help reconstruct patterns of human migration around the globe.  Subscribe to this podcast. Plus, to stay updated on all things science, sign up for Science Friday's newsletters.

This is Science Friday. I'm Ira Flato coming to you from the Sydney Goldstein Theater in San Francisco. One of my favorite topics to discuss on Science Friday is the microbiome, you know, that vast menagerie of trillions of microscopic bacteria and fungi that live on us and inside of us. But I found out recently there's a lot more living on us than just microbes. Take, for example, the face mite. Yes, you heard me correctly. It's a tiny arachnidna. that lives on your face. And in some studies, it's been found in 100% of people sampled. So chances are yes, they're hiding out in your oil glands and hair follicles right now. I see a few of you squirming in the seats about that.

Yeah, I understand it. And you know, if you do the math, there are about 5 million hair follicles on your body, and there are more than 7 billion people on earth. So there are lots of face mites everywhere you look. And joining me now to talk about these ubiquitous creatures is Michelle Troutwein, curator of entomology at the California Academy of Sciences here in San Francisco. Welcome to Science Friday. Thank you.

Okay. What exactly is a face mite then? So a facelite is an arachnid. So it's related to a spider. They're microscopic. And they kind of look like a stubby little wets. worm with eight little legs at the tip. And it's safe to assume that everybody in our audience might have one at least? At least one. I hate to break it to you, but yes, you all have face mites, probably

dozens, hundreds, maybe thousands. Ooh, doesn't that make you feel good? But we call them face mites, but they live all over your whole body. All over your body. They really like the greasy spots, so they congregate mostly in your face, but also your genitals, your ears, nipples. nose inside your nose wow and I'm afraid to ask the next question what do they eat on my body so they eat sebum which is kind of just the oily gunk

on your skin so that's why they like the greasy parts because that's what they eat although to be honest that's what we assume we know very very little about them so maybe they're eating bacteria maybe they're eating skin cells there are still a mystery okay so presuming that the oil

that my pores are producing is meant to protect my face from the elements, right? So having all these arachnids eat it, isn't that depriving me of my own facial oils that I need? It doesn't seem like it, right? It seems like we probably have enough. And maybe even they initiate the creation of sebum. I don't really know. And we pass them from one person to another? So that's right. We all have our own population, and it seems like they're passed mostly from parent to offspring. So we do, I have some data that shows we share them between people you're really physically close with, so you're, you know, your partners, but it seems they're primarily passed down from parent to offspring.

So what if you try to wash your face enough and you use all kinds of creams and lotions or whatever, you're shaking your head like, I'm wasting my time. You're wasting your time. Yeah, so I don't think washing does anything. If you really put some kind of pesticide on your face, you could, no, really, you could kill some of them, but here's the thing. Don't try this at home. Don't try that at home. Here's the thing. They also can reproduce asexually. So if there's one female left anywhere, the whole thing will start all over again. So they're just a part of your skin ecosystem and an inevitable part of our life on Earth. If they don't poop, right? I understand they don't poop. What happens to them? They're going to blow up?

When we first learned that, it seemed crazy. But actually, it turns out plenty of mites don't have an anus. That's actually not. as unusual as it sounds. So mites, some mites, just excrete their waste in other ways. Wow. So, you know, whenever we talk about biology and animals and humans, we talk about they occupy some niche in nature, right, a niche. What is that that the face mites? Why do we have face mites? Where do they fit in? Yeah, they've created this amazing niche for themselves somehow in early mammalian evolution. So this genus, originated with mammals, and as every mammal species arose, mites arose with them. And actually, we have two species on us, and they live in separate microhabitats on our skin.

So the niche even becomes an even smaller niche. Wow, you sent me a facemite kit, you called it, where I actually dabbed oil on my face, and I scraped it with a spoon, and I daubed it in some cotton, and sent it back to you, so you could look at my face mites. What did you learn from that? Yeah, so one of the most interesting things about face mites is that different people from different parts of the world host different lineages of face mites.

So I can tell you something about your ancestry by looking at your face mites. Are you kidding? No, so... I'm afraid to ask the next question. So surprise, surprise, this is the big reveal. Your face mites fall into the European lineage of face mites. You didn't need face mites to know

that. Let me tell you that, but that's good to find out. And so how did they decide they wanted to live in our pores? This is an evolutionary thing that's gone on? That's right. So mites live in the craziest places all over in every aspect of the world you could ever imagine. So there's mites that live in the noses of birds and then the feathers of birds and then the skin of frog. So, you know, our bodies are just habitat, right? And these mites have found great, cozy places to live. about face mites and you'd like to ask a question up here. We have a couple of microphones. I'm wondering if other animals, mammals, also have face mites. Yes, so all mammal species have face mites. And what's amazing is that right now there's about 150 described species in this genus.

But we know almost every mammal that we've looked for has them and some have two, three, four species unique to them, which means there's thousands and thousands of face mites to be discovered on mammal species all over the planet. So we need some new scientists, some young budding scientists to come and study face and mites with me. So am I swapping mites with my pets? Well, if you are cozy with your pets, you could get one of their mites, but they're so uniquely evolved to each host that they probably won't do so well. So I wouldn't worry about getting taken over by your dog mites or cat mites. Well, good to hear that. I think we have a question. Gentlemen, coming to the microphone right here.

Yes. My hair is starting to fall out. Are they the cause? So, believe it or not, face mites really like fine hair, so there's not a ton of them on your scalp. They really like the fine facial hair and hair in other parts of your body.

So you can't blame them for your hair loss, no. But you're giving them extra space. If your hair is gone. That's true. That's true. They're probably going to move in. You're helping me. face mit population. Well, I understand that you study face mites on whales, too, right?

Well, this is a big mystery. Face mites have never been described on whales. But how crazy would it be if whales had face mites? So at the California Academy of Sciences, some of our team often responds to the whale strandings. And so I have been delivered a gift of a big chunk of whale skin with a hair follicle and even a tiny hair. And I looked for whale mites and I have to say I didn't find any, which was tragic, but one day, you know. That's how Captain Ahab got started, right? Let's go up there in the balcony, yes. Are skin rashous like eczema related to mites?

Not eczema, but they are associated with rosacea. And, you know, sometimes I think they get a bad rap. It's not clear if they cause rosacea or if they're just a lot of them, if they're just correlative with rosacea. Have you genes sequenced face mites? Yeah. So I've been working for years to sequence the face my genome, and it is so problematic because it's such a tiny creature, and it's full of bacteria, and it's full of human DNA. So what I've done is sequenced, you know, my genome and my postdoc's genome over and over again.

But I haven't gotten enough face my DNA to put together a genome yet. I'm working on it. Now, we were told that parents can pass on their face mites to their kids. right, in breastfeeding or just kissing them on their forehead? That's right. Like that? Does one parent usually do it more than the other, the passing on? Well, so, you know, I think traditionally they were probably passed down from mother to offspring

primarily through breastfeeding. Although I can say in my own family, I've tested my husband and my kids' face mites. Wait a minute, wait a minute. I don't know. Hey, kids, we're going to test your face mites today. Yes, I did. Right after breakfast. They didn't have a choice.

And I can say that my kids have my husband's face mites, which, you know, after all the time I put in, you know, it really bummed me out. Wow. I must have really hurt. Yeah, it did. So when the Europeans do all this, all this cheek kissing all the time, you know, they do two cheeks instead of just one,

they're passing a lot of face mites. I don't think so. I think it's got to be a little bit more prolonged contact. Otherwise, so what I found is that I've sampled thousands of people and everyone has unique face mites unless you're together in a family, right? But you don't share face mites with strangers. So if you were passing them along every time you kiss cheeks, there would be more sharing amongst the population. And there's really not. Your face mites are really your own. I'm going to remember, I'm going to write, you know, got a little pincushion. Your face mites are really your own? Yes, over here. It's romantic. Are there any forensic uses for face mites? You know, there would be forensic uses. Now, I don't know if anyone is, is, is, you know,

using them, but it would be, it's a, it's a field ripe for, uh, ripe for forensic use. Um, for example, probably because your face mites are really your own, uh, you can tell, uh, you know, you can leave evidence behind. You can leave evidence. That's right. You can leave your face bites on them. I've always thought it would be a good sort of like, uh, infidelity test, you know, if you, you can, you have, check your partner's face mites. You have the makings of a TV writer. I can, I can tell you that. Thank you very much for, yeah. taking time to be with us today. Thank you.

Thank you. Michelle Trout Wine, curator of entomology at the California Academy of Sciences here in San Francisco. Thank you very much. And if you're just itching, see what I did there,

for more might knowledge, you can sign up to receive a might text every day this week by texting the word might to 917-242-4070. That's mite, M-I-T-E, to the number 911 2424070. You get a daily mite delivered to you via text.

It's not a real mite. No, no, it's not a real face mite. We're going to send you some information about that mic. And please welcome our musical guests for the evening, the Bay Area's own Fox Tales Brigade. This is Science Friday. I'm Ira Plato. Science fiction can take us to worlds that feel very different,

whether it's the robot uprising, a galaxy far, far away, or even as they say, wibbly, wobbly, timely stuff. But these stories only appeared to go, as they say, where no person has gone before. As far-fetched as tales of the future may feel, my next guest say they can actually connect us to familiar quandaries. Take Star Trek's Commander Data, a super smart android crew member in the next generation. In the season two episode, the measure of a man, data is suddenly on trial.

Is he property or is he a person with the right to choose what happens to him? And at the heart of the matter, is he sentient and self-aware? Here's a clip of Captain Picard trying to test just that question. Commander Data, what are you doing now? I'm taking part in a legal hearing to determine my rights and status. Am I a person or property? And what's at stake? My right to choose.

Perhaps my very life. My rights. My status, my right to choose. My life seems reasonably self-aware to me, commander. Now let me welcome my guests. Annalie Newitz, the author of The Future of... another timeline. Charlie Jane Anders, the author of The City in the Middle of the Night. I know that you both are co-hosts of a podcast called Our Opinions Are Correct. That is correct. That is as correct

as our opinions. On Science Friday, we love to talk about taking STEM and turning it into steam. You put the A from Arts into STEM, and this seems to be a perfect example of how you take the arts and talk about contemporary issues, right? That's right, and I think of science fiction as kind of the cultural wing of the scientific project. I mean, part of what our job is as science fiction writers is to spin up scenarios that people who are working in labs or who are slaving away at a keyboard

don't really necessarily have time to think about. But what I think is really great about this Star Trek episode is that it raises all of these possibilities that you might not consider if you're developing AI. It also raises a lot of ethical questions about how we should treat data, but also about how we engage in the scientific process

because one of the big questions is, will a scientist be able to vivisect data? So Charlie Jane, how does the episode resolve all of these ethical issues? Basically, it's resolved by everybody deciding that data actually is a person rather than a thing, and the way that that's kind of established in part is because the scientist who wants to dissect data, Commander Maddox,

has been using the pronoun it to refer to data throughout the episode. But at the end, he finally uses the male pronoun to talk about data. And watching this episode recently, I was really moved by it because as a transgender person, the fact that data's pronoun is finally acknowledged felt really meaningful to me. And it felt like that's part of acknowledging his personhood. And it's also acknowledging that his experiences and his, sense of himself are valid. And one of the things that they keep talking about in the episode is whether

data's memories or just information that you could back up to any other computer, or whether there are lived experiences and signs of a journey that he's been on as a person. One of the things that's interesting is that this is an episode that really deals with robots and legal procedures. But we've actually just had one of the very first trials in the U.S. over the Uber self-driving car and who would be responsible when it killed somebody, when it ran over someone. And this is a question that revolves around what responsibilities we bear for the automation that we're unleashing on the streets. And the other thing that's really relevant, I think, about this episode is it deals with the

way that bias creeps into the scientific process. Because Maddox, who is the scientist that wants to vivisect data, what he thinks when he sees an artificial intelligence is, well, why don't we build an army of disposable labor? So he looks at the AI and he doesn't think, hey, this could be a pal or this could be a colleague. He thinks, aha, I found a disposable laborer. Excellent. How can I make thousands of them? Or millions. Yeah, and in fact, that's always been part of the stories about robots. And in this episode, there's an amazing moment where Geinen, the kind of bartender slash confidant, talks to Captain Picard about the higher implications of allowing data to become property.

Well, consider that in the history of many worlds, there have always been disposable creatures. They do the dirty work. They do the work that no one else wants to do because it's too difficult or too hazardous. And an army of data is all disposable. You don't have to think about their welfare. You don't think about how they feel. Whole generations of disposable. people. That's a little harsh. I think that's the truth. That's a truth that we have obscured behind

a comfortable, easy euphemism. I mean, how often does sci-fi like the Star Trek quandary get you thinking about contemporary technology and society issues? I think that as a science fiction writer, I'm always thinking about how I can create scenarios that might help to inform science, but also I'm informed by science. I'm constantly thinking about science and kind of curious about how things could develop based on different technological or social changes. And I think that, you know, people talk about thought experiments in science fiction. I think it is, obviously, it's not a controlled environment. I mean, my brain is probably the least controlled environment you could possibly imagine, but it is a way to sort of pose these what-f scenarios and kind of game them out. Yeah, I mean, I think that,

you know, especially around something like artificial intelligence, which has been kind of a dream for humans for decades at this point, it's still really in the realm of science fiction. And if we're going to be developing living beings that are human equivalent, we have to bring in these issues around ethics and culture. Yeah, and I think that there's no limit to the number of issues in society and in science that you can point to instances of them in science fiction. I think that once you start looking, they're everywhere. and a lot of stuff that we kind of think of as ubiquitous in our world

came out of science fiction one way or the other, from our gadgets to the idea of sciops and propaganda, which we talk about in our podcast, and then some of the environmental issues that science fiction deals with, which are now increasingly pressing and kind of a part of our lives. Yeah, one of the things that we talk about a lot is the fact that if you're going to be imagining the future now, you absolutely have to imagine climate change and climate crisis, because otherwise you're just not presenting a realistic picture

of what tomorrow will look like. You're both, besides prolific podcasters, you're also both authors. So let's talk for a second about your recent books, Annali. The Future of Another Timeline is a story about time travel, right? That's right. And I was actually very disappointed when I started working on it when I interviewed a couple of physicists

and found out that actually time travel doesn't, and will never exist. So it turns out that that is not very scientific, but I wrote a book about how scientists try to grapple with it and understand it when they do discover time machines, basically embedded in really ancient rock formations on Earth. And they understand that they don't seem to fit with our models of physics, but they act very much

like geophysicists in the way they go about analyzing it. So it was really fun for me to have a book where I got to give you a little bit of fantasy, but also include the scientific process. You say that the story was inspired by something in the present, my stuff that we're doing in the present. It's true. It was inspired by Wikipedia Edit Wars. And that is because my characters who go back in time and think about changing their

timeline refer to it as editing the timeline. And eventually my characters get into an edit war with another group that keeps reverting all of the changes that they're making. So my heroes are feminist time travelers who are going back in time and trying to improve women's access to reproductive health care. And there's these guys, we need it. There's these guys that keep going back

and reverting their edits. And basically they wind up in the year 2022, which is their present day, in a United States where there are no abortion rights for women. So they have to fight this edit war. It's the same edit war. We've all fought online and all fought on Wikipedia, but they're actually finding it over the timeline.

And it's quite difficult to change the past. It's not easy at all. You really have to go back and spend years organizing. So I had a lot of fun making it feel like actual scientific work where you don't just instantly find something. You actually have to do assays for like 20 years before you actually figure out how face mites work. So I think that it's really, that's part of the fun is getting the audience to feel like, yeah, we're involved in science and an edit war.

Let's move on to your book, Charlie Jane. The City in the Middle of the Night tells about a world 3,000 years in the future of humanity, and you put it on a barely habitable planet. The City in the Middle of the Night is a book that entirely came out of a science obsession of mine. while we were working on I-O-9, I became obsessed with tidily locked planets, which are planets where one side of the planet always faces the sun, and one side always faces away from the sun, the same way that one side of the moon always faces Earth,

and there's a side of the moon that we never see. And on a tidily-locked planet, basically, if you're living there, the sky is never going to change over your heads. There's no sunrise, there's no sunset. And, you know, depending on who you talk to, the only part of the planet that might be habitable to humans is this thin strip of land called The Terminator in between the dayside and the night side. And so I actually thought about calling my book The Terminator, but it turned out that that title was taken.

I don't know. I don't know why. And in your story, human technology is actually failing, right? The colonies collapsing. Why did it seem like that was an important feature to explore? Yeah. I mean, in general, I spent a few years. is really trying to imagine what it would be like to colonize a planet like that.

And every aspect of it, like, how would you deal with the passage of time and knowing when to sleep and understanding how much time had passed when you don't have these cues of like the sun going up and down? And just how do you organize your society and how do you manage your resources? And one of the things that I decided over a period of time when I was thinking about this is that eventually our technology would start to fail because after hundreds of years, I mean, they don't know how many years it's been, but after hundreds of years on this planet, there are things that you can't replace anymore.

There are things that it's hard to make with the facilities that they have. It's hard to keep building computers. It's hard to keep building all these really high-tech things that they came to the planet with, and they kind of end up with, like, kind of uneven technology where some things are still pretty advanced, and some things are just not working anymore. But also, in general, it's getting harder and harder to live on this planet because we've been messing around with the environment, but also it's just a challenging environment place to live.

It sounds like our planet. Is that what you're talking about climate change? And if you're thinking that technology is going to save us, it might not save us? I mean, that's certainly part of it. I think that, you know, everything I write, climate change is in the back of my mind because it's such a huge part of our lives now. And, you know, living in San Francisco where sometimes you can't go outside without a mask because of the smoke from the fires and, you know, hearing about all the natural disasters

all over the place. It's hard not to be thinking about climate change and about how humans interact with our environment and how we do tend to kind of, you know, have an impact on our surroundings, both good and bad, and that we need to be more careful about how we approach, you know, an ecosystem that, especially one that is not our ecosystem to begin with, which is part of what I was interested in here. I'm Ira Flato. This is Science Friday from WNYC Studios. Yeah, you can applaud if you like. We're here in San Francisco talking with authors Charlie Jane Anders and Annalie Newitz. Also, they co-host our opinions are correct podcast.

Annalie, what are some of the themes? And if you read science fiction, you see themes coming up over and over again for you. So I would say the themes that I seek out in the kind of reading that I do and in the writing I do tend to be around how, people resist when they're in a situation that is seemingly unsolvable. So it might be an environmental system where people are struggling with climate change. But a lot of times I deal with social systems and how people try to form groups that can resist. And this is a huge theme in science fiction. It's everything from robot uprisings, which let's remember, I mean, robots have uprisings because they're being oppressed, all right? It's not just because they want to

eat humans or bomb us. It's because we turn them into slaves and they're pissed off. Nobody likes being a slave, including a robot. So I find that really interesting. I like kind of finding the chinks in a system. And so whether that's a futuristic corporation or a planet where you're trying to terraform, I just, I really love seeing people connect with each other, even if it's not humans, it can be aliens or robots or ants. I just like them getting together and, yeah, defeating some kind of faceless enemy. You're smiling, Charlie Jane. What do you think are some of the recurring? I mean, I'm very interested in reconciling opposites. I'm interested in people who are caught between apparent, like complete contradictions, like two completely opposite things, and how they deal with that.

And in this book, it's about people who are caught between the day and the night, and these are two kind of opposite sides of the planet that they're living in between. But I'm also really interested in empathy. I think that my work always tends to lead towards people having more empathy for each other and for people that they don't even consider people and learning to just have more understanding. And I feel like this is a thing that I would love to see humans do in real life, having more empathy for the other, having more understanding of people who aren't like us, having more openness to other ideas. So that's something that I can. keep kind of chipping away out in my work.

I think both of us are really interested in writing hopeful stories too. And I think this is a bigger movement within the science fiction and fantasy community is to think about stories that aren't dystopias or that are difficult dystopian worlds where people are figuring out ways to change it or reform it or survive it. Because sometimes just the act of surviving can be incredibly hopeful in dark times. Yeah, I think it's really important to offer people like a message of hope and not just like easy hope, but hope through struggle, hope through doing everything that we can to make things better. We're talking about the future of technology and how science fiction can inform the real world. When we come back, is our tech future a dystopia or something more hopeful how science fiction and ethicists think about what's possible?

Taking us to the break, our music. guest, Fox Tales Brigade. This is Science Friday. I'm Ira Flato. We're talking about the future of technology and how science fiction can inform the real world with my guest, Annalie Newitz, the author of The Future of Another Timeline,

and co-host of the podcast, Our Opinions Are Correct. Charlie Jane Anders, the author of The City in the Middle of the Night, also co-hosts, our opinions are correct. I want to take us more into the real, world now and the real future, especially where AI is concerned. This is after all the Bay Area and it's home of countless companies contemplating how to make tech and artificial intelligence specifically do more of the heavy lifting in our

society, social media, ride share apps, even AI to read x-rays, and so much more. But as recent headlines have highlighted, it's not as easy as writing an algorithm to to curate our news feeds, help us find a ride downtown, or decide if we need surgery, AI makes mistakes. It has unintended consequences, or does exactly what it's designed for but to the dismay of some portion of society. Who's out there thinking about the most responsible way to develop tech? Well, my next guest is let me welcome to this stage, Dr. Ruman Chowdhoutary, a data scientist

and lead for responsible AI at Accenture Applied Intelligence. Welcome to Science Friday. Thanks for having me. Dr. Chattery, when we think about AI in the future, there are a couple of common flavors. We have the good guys, the robot butlers, and the bad guys. We have the AI apocalypse, Skynet. Are you thinking about either of these as possible outcomes from the AI we have today?

So that's a pretty complex question, and I love the fact that you've led with two science fiction authors, is that there's so much of what we build, which is limited by what we are told and what we can imagine. So even the images you've put up there are usually from a dystopian future. And what we need to have are very positive narratives or different ways of thinking about artificial intelligence.

And that's kind of what I do, even though I'm a consultant. I work with companies to help them think of how to use artificial intelligence in ways that are beneficial to humanity. And how do you think about that? I mean, how do you guide people to be responsible with AI? Well, there's a lot of grounding in reality. So actually, I talk more often about algorithmic systems rather than artificial intelligence. So AI often evokes this idea of magic or things that are too big or too heavy for us to understand.

An algorithmic system is about how an algorithm, which is essentially math put into code, interacts with the real world. And the context in which we live today and probably more importantly, the historical context upon which it was built. Do you have to first come up with a definition of what responsible AI is? Absolutely. And, you know, when I think about companies, often these things are already enumerated in their core values or the mission statements. So they'll say things like, you know, we have a dedication to diversity. Well, if you're building an artificial intelligence or an algorithm to help you in your hiring system, that's going to directly impact your ability to have a diverse workforce.

Can you actually train people to create more responsible AI? Absolutely. That's my job. I couldn't do it all myself. There are a lot of folks interested in how we build ethics into the development of artificial intelligence systems. There are a lot of social scientists. So my background, I'm a statistician of social scientists. I'm not a computer science major. And I think that's one of the most beneficial things that people like myself bring to the field is an understanding of humanity and social sciences. Can you give me an example of where we might see today AI being irresponsible? I think that often things are built without thinking of what I call question zero.

Just because we can doesn't mean we should question. So there are some papers on how much energy it takes to run a complex AI model when we're thinking about climate change. So recently there was a natural language processing. So like a language generation algorithm built that was so convincing the people who created it were worried that people people could just use it to create fake news. Or another example was deep fakes, which is a technology that's publicly available, was used to create an app called Deep Nudes, which could basically take an image of any woman and regenerate her naked.

So there is plenty of irresponsible AI being built out there. And one of the things that a lot of companies, a lot of organizations are thinking about, is something called responsible role at a responsible release. So if we're creating an algorithm that could be used for both, good and bad, how do we roll it out or protect us against the bad uses while still allowing for the good uses? Yeah, we have done some stories recently about systemic bias and algorithms, for example, used by law enforcement, facial recognition databases, things like that. Is it as simple as fixing training data? That's one part of it. It's not all of it. So facial recognition is a great example.

Some of the earliest work was about how facial recognition doesn't recognize darker skin and particularly women of darker skin color. But the question really is, should facial recognition exist in a lot of places are banning it? So, you know, is the answer to make better detection materials for law enforcement to oppress minority communities? No. It's not just about perfecting it. It's about understanding about how it's being used. When we ask humans to participate in the criminal justice system, they're going to have biases. So why is the bias of an algorithm in predictive policing worse than what people might have?

You're absolutely correct. So actually just wrote a piece for VentureBeat on exactly this. It's something I call the retrofit human. When we build technological systems, they are limited, and we sort of shove humanity to fit the limitations of these systems, right? So when we think about, for example, self-driving cars, I think we all imagined the same thing when, you know, it was told to us like five or so years ago. We imagined like living an hour outside of where we worked, getting in our car napping while our cars whisked us off.

And now what we're seeing, and, you know, the Uber example is a great one. We're going to have to, and Tesla too, we're going to have to sit in our cars with our hands at 10 and 2 and our foot on the brake, paying attention to the road. Well, what benefit did this bring us then? The technology is quite limited. So what we've done is there's this promise, and often an over promise, then there's this hard reality,

and we're still kind of having to fit human beings into this flawed system. So in criminal justice, it's a great example, because you have judges being given the output of these criminal risk assessments. Now the question I ask is, do we feel empowered as human beings to question the algorithm? And do we know enough about it that we can dissect the outcome

and understand if there's something good or bad happening. All right, we have a lot of questions in the audience. Let's go up to the balcony, yes. Hi there. I'm curious. Star Trek also had an episode in which the holographic doctor created a novel, and I'm curious, both where do you think AI has ownership over the things that it creates, and if that case were decided today, where do you think it would fall in our justice system?

The novel you refer to is photons be free. And it is interesting because that is set in the future after we've already established that data is a person. And yet we're having to ask those questions again about the hologram. And I think that that really points out how much we think of our future AI companions as potential servants. And I think I'm really curious when we think about building ethical AI, is that something that scientists? are thinking about, about the kind of assumptions that we're making about, you know, what exactly these human equivalent beings will be doing for us. I mean, I think there is certainly a camp that thinks pretty far ahead.

I mean, what I will say to maybe address the question, GPT2, which was the language generation algorithm I was talking about, essentially can write stories. They don't necessarily, they're not necessarily particularly good, but they're realistic enough. I think somebody had generated an essay that was submitted to some magazine competition, I think possibly at the economist. And, you know, it wasn't bad. It kind of, you know, having graded enough papers in my life, like an average high school to undergrad student paper, which is not bad for an AI. When you think about things like ownership, though, that's really interesting. And this is where the legalities come in.

Here's a problem with giving an AI ownership. then it would also have liability, right? So if something bad happened, what do you do? Are you sending an AI to jail? Do you turn it off? It doesn't feel like we're not, we haven't reached a singularity. Do you think we're going to be having a singularity? No.

No. There's not going to be a time where AI gets smarter than people is what I'm asking. And because there were a lot of smart people who are warning us about that. There are. And there are also a lot of not-so-smart people who talk too much about it. Well, and the reason I say that is because sometimes it distracts from the very real issues where, you know, there was a human being or a corporation that created this AI. And frankly, it can absolve them of the liability of the product that they maybe have built. If we're shoving, this is actually another term I call moral outsourcing.

We over anthropomorphize the limited AI that exists today. And we shove off the responsibility and we say, well, I don't know, you know, I'm just the engineer. I built it. but the AI is, quote, making its own decisions. And that's true to an extent, but as somebody who's designed these systems, we're setting optimization parameters. We're not telling it exactly what to do, where we're kind of telling it how to do it. I would just want to add about the singularity, too, when we talk about AI becoming smarter

than humans, we still don't have a definition of intelligence for humans. We don't really know what it means to be intelligent. We do have a sense that there's certain people who think they know what it means to be intelligent, but we also know that there's many ways of being conscious. There's people who are neurotypical. There's people who are non-neurotypical in all kinds of ways. And so I don't think we can with confidence ever say that we're going to make something smarter than people because we don't yet know how smart people are.

And to your point, like, you know, computers have faster processing power. They're never going to get tired. They're already kind of smarter than us in many ways. Sure. I would just add that we are starting to see more creative work done by AIs. There was a great article in Mother Jones a while ago about A.I.'s writing music and how AIs can now write melodies that are as good as, you know, Ed Shearin, which, you know, that's a very low bar. But eventually they might get to Katie Perry. I don't know. But, you know, I think that what the real test is, when will there be a time when an AI writes a song or a book and is upset that it doesn't get credit?

Like, that's the real test for me. That's the real litmus test is, when will there be an AI that's like, no, I wrote this. And the only other thought I had is that Ted Chang, who's an amazing science fiction writer, has been saying that a lot of our fears around AI's and, like, domination by evil supercomputers are actually our fears of corporations, like that we're just projecting that onto this mythical AI. I'm Ira Plato, and this is Science Friday from WNYC Studios. So to your point about AI writing music or writing novels, that's pretty much a theme. So I was teaching data science before I joined Accenture, and my students had done projects,

like, can I create an AI that replicates classical music when my students had done this? And pretty much he created an AI that created more bocky-bock, you know, or more Mozart-y-Modart. It can kind of just copy what's been done, maybe tweak it a little. But, you know, I think it would be truly novel for it to write something that had never been written or, you know, that you couldn't really understand the origins of. And that's something human beings do, not something we've been able to do with AI. Well, let's see if we can get one last question here. We're seeing a lot of our technology use more and more rare earth metals, lithium, coal. What is the eventual end goal for this?

What company will control these rare earth minerals? And what if it is the military? What if it is the mega corporations of the cyrumpur dystopian dystopia? So I think in today's narrative, we absolutely do need to think about power structures. And again, that's the social scientist in me speaking. So I absolutely do agree with the sentiment. We do live in a world in which there are the haves and the have-nots, and that's actually increasingly getting more and more divided.

I actually was thinking about the Geinen sentiment from earlier, the clip you were playing, and I was thinking about two books that I would recommend anybody who's interested in this topic reading. One is Sarah Roberts' book on content moderators called Behind the Screen, and the other is by Mary Gray and Siddhar Tzu, called Ghost Work. And both of those are about the hidden labor behind artificial intelligence. And things that may seem like magic are actually things. things, and we get further and further removed from this labor. There's very cheap labor or individuals in other countries that do these tasks of coding

images so we can have computer vision or watching terrible videos on social media so your child doesn't have to be exposed to it. Those two books I would highly recommend reading to really understand the power dynamics behind artificial intelligence, and actually we're the 1%. We're the ones that receive the benefits. Most of us probably live in San Francisco or somewhere in the Bay Area. We get all the nice things, and we don't have to see the ugly side of sometimes how this technology is made. Yeah, and a lot of times when companies, like, say, Facebook or Google, talk about automating a lot of aspects of their search or of their moderation.

They actually are using people for that. And they might be using little bits of automation here and there, but it would be. really is the case there's a ton of human labor, and a lot of it is task work. It's people who are working by the minute or by the five minutes. They're not unionized. They have no opportunity to organize with each other because they often don't see each other. They just meet in forums or they get their work through various portals online. And these are the people who we think are actually AI. You know, these are the people who are making sure that when you do a Google search that you actually get an accurate result or a result that resembles what you'd like to see.

And I think that's one of the great lessons of thinking about robot uprisings, because of course we all know the robot uprising is really the worker uprising. It's really about people who are doing labor and being deprived of access to resources who just get sick of it. And so I think it's so fascinating that we're living at this moment when we hide human labor behind the idea of automation. And then at the same time, people feel threatened that automation is taking their jobs, even though it's really just people who are really underpaid who are taking your jobs. It's not actually automation quite yet. Well, terrific discussion. I want to thank my guest, Annalie Newitz, author of The Future of Another Timeline,

co-host of the podcast, Our Opinions Are Correct. Charlie Jane Anders, author of The City in the Middle of the Night, and also co-host of Our Opinions Are Correct. And Raman Chattery, a data scientist and lead for Responsible AI at Accenture. Thank you, for coming by and taking time to deal with us. That's about all the time we have. Our heartfelt thanks to Jonathan Blakely,

Ryan Davis, Janet Lim Young, Sevda Eris, Lawrence Summer, Josh Cassidy, and all the great folks at KQED for hosting us. And we also want to thank Ray Livingston and to everyone at Another Planet Entertainment and to all the amazing staff here at the Sydney Goldstein Theater. Thank you for helping us out. for helping us out.

Oh, and let's give one last round of applause for Fox Tales Brigade who will play us out tonight. Thank you for coming. San Francisco, I'm Ira Plato. Drive safely, everybody, and good night.