Latest / kill switch / AI & Us
Transcript
- 0:04Speaker 1: Sleepwalkers is a production of our Heart Radio and Unusual Productions. Hi,
- 0:17Speaker 1: I'm Aloan and I'm care Price. Welcome to a special
- 0:20Speaker 1: bonus episode of Sleepwalkers. Well, Cara, it's very good to
- 0:36Speaker 1: be back to the office. That you It is good
- 0:39Speaker 1: to be back, as I just want to apologize to
- 0:41Speaker 1: our listeners. We don't have an algorithm that's going to
- 0:43Speaker 1: create season two, so things have taken a little longer. Yes, unfortunately,
- 0:48Speaker 1: unfortunately a I can't do everything yet. But we are
- 0:51Speaker 1: hard at work on season two of Sleepwalkers, and we're
- 0:54Speaker 1: focusing on stories that really contextualize the implications of AI,
- 0:57Speaker 1: what it's doing, what the future is, and how it's
- 1:00Speaker 1: affecting us. You know, we had such a good time
- 1:03Speaker 1: in season one of Sleepwalkers wrapping our heads around the
- 1:06Speaker 1: meaning of artificial intelligence. It becomes basically a principle of
- 1:10Speaker 1: statistics prediction, how we're using data to inform our decisions,
- 1:14Speaker 1: and how that's becoming ingrained in products and services and
- 1:18Speaker 1: everything we do. Really, it's true the Pandora's box of
- 1:22Speaker 1: AI has been opened, but we still have the black
- 1:24Speaker 1: box problem that is true explainable AI. We can't tell
- 1:27Speaker 1: what neural networks are doing yet, but people are working
- 1:30Speaker 1: on it, and that's a story we're going to be
- 1:32Speaker 1: covering closely in season two. But in this bonus episode
- 1:35Speaker 1: of Sleepwalkers, we're going to take a look back at
- 1:37Speaker 1: some of the most poignant stories and interesting applications of
- 1:40Speaker 1: AI that we talked about in the first season, and
- 1:43Speaker 1: later in this episode, I'm going to give you a
- 1:45Speaker 1: preview of a fascinating new company that's using AI to
- 1:48Speaker 1: predict very specific consumer taste, as in preferences, not like
- 1:54Speaker 1: tasting clothing. Before we get through though us, when you
- 1:57Speaker 1: look back at season one, what stands out to you
- 1:59Speaker 1: the most? One of the things that stands out to
- 2:01Speaker 1: me most is the story you did about liar Bird.
- 2:04Speaker 1: Thank you, but seriously, the way they use an algorithm
- 2:07Speaker 1: to create a deep fake of your voice. But in
- 2:09Speaker 1: that particular piece, the questions it raised about mortality and
- 2:13Speaker 1: would you want to hear your father's voice beyond the
- 2:16Speaker 1: grave has stuck with me and I was one of
- 2:19Speaker 1: the most powerful moments in the whole podcast. Yeah so,
- 2:22Speaker 1: Jose Satello, one of the founders of Liar Bird, had
- 2:25Speaker 1: me sit down on a microphone for an hour and
- 2:27Speaker 1: just speak which was a personal dream of mine, and
- 2:30Speaker 1: then using that out me interrupting. Using that voice data,
- 2:36Speaker 1: liar Bird's algorithms created a version of my voice that
- 2:39Speaker 1: could turn any written text into something that sounded like me.
- 2:42Speaker 1: But Jose actually explained it better. So can you hit
- 2:46Speaker 1: the clip? Not an AI scientist, but we do have
- 2:49Speaker 1: the sophistication to roll tape. I know it might sound
- 2:53Speaker 1: a bit like magic, but in reality, the way that
- 2:57Speaker 1: our algorithm's work is basically they are just a battern
- 3:00Speaker 1: matching algorithms, and so it's trying to figure out how
- 3:05Speaker 1: to identify the patterns in your voice by comparing it
- 3:09Speaker 1: against thousands of other voices a shoually tens of thousands
- 3:12Speaker 1: of other voices, and trying to figure out what is
- 3:16Speaker 1: it that makes your voice unique. Once Jose's algorithms identified
- 3:19Speaker 1: what was unique about my voice, obviously everything they had
- 3:22Speaker 1: the building blocks they needed to make a fake. Then
- 3:25Speaker 1: we sent Jose a set of sentences we wanted robot
- 3:28Speaker 1: care to say, and he used another set of algorithms
- 3:31Speaker 1: to turn the text into what we heard. The way
- 3:33Speaker 1: they do this is they use it's called a generative
- 3:35Speaker 1: adversarial network again, which is a system where one neural
- 3:40Speaker 1: net tries to trick another one a thousand times per second,
- 3:44Speaker 1: So each time the second network to texts of fake,
- 3:47Speaker 1: the first one tries again. It basically learns from its mistakes,
- 3:51Speaker 1: and once it tricks its adversary, it's ready to show
- 3:53Speaker 1: its results. In our case, liar bird pits my fake
- 3:57Speaker 1: voice against my real voice until it sounds like this
- 4:00Speaker 1: sub dog Scara Karen. One of the reasons why this
- 4:04Speaker 1: story has stuck with me is because it feels like
- 4:07Speaker 1: we're just at the beginning of tapping the potential and
- 4:10Speaker 1: the potential for harm of deep fakes. May be remembered
- 4:14Speaker 1: as the year of the first significant deep fake crime.
- 4:17Speaker 1: An employee at a UK based energy firm believed he
- 4:21Speaker 1: was on the phone to his boss and followed instructions
- 4:23Speaker 1: to transfer two hundred thousand pounds to a scammers bank account.
- 4:28Speaker 1: That certainly won't be the last deep fatefore we hear of,
- 4:30Speaker 1: and it raises questions about responsibility and accountability. Who's liable
- 4:35Speaker 1: in a case like this. Facebook has actually gone so
- 4:38Speaker 1: far as to create a deep fake detection challenge to
- 4:41Speaker 1: get the best minds thinking about deep fakes and how
- 4:44Speaker 1: we might solve the problem and offering like a million
- 4:46Speaker 1: dollar prize reach prize, which is like a dollar but
- 4:50Speaker 1: it also shows how important the issue is, especially when
- 4:52Speaker 1: a company like Facebook gets behind it. Um. There's another
- 4:55Speaker 1: side to deep fake technology that actually highlights this dichotomy
- 4:59Speaker 1: and tech anology right now, which is that it can
- 5:01Speaker 1: be used for menacing purposes but also really powerful and
- 5:05Speaker 1: beautiful applications. Jose goes on to talk about how liar
- 5:08Speaker 1: Bird can be used with als patients and give them
- 5:11Speaker 1: the ability to speak when they've lost all ability to speak,
- 5:14Speaker 1: when it could give them the opportunity to speak in
- 5:17Speaker 1: a version of their voice to their children again, which
- 5:19Speaker 1: is quite profound. One area where I think technology is
- 5:23Speaker 1: a powerhouse for change is in medicine. Technologists and doctors
- 5:26Speaker 1: alike are looking at AI to predict, treat, and diagnose,
- 5:30Speaker 1: you know, everything from depression to cancer, and that's a
- 5:33Speaker 1: very wide spectrum. And it reminds me of one of
- 5:36Speaker 1: my favorite interviews that you did, which was your interview
- 5:39Speaker 1: with Saddartha. Muker g just so fascinated by this article
- 5:43Speaker 1: he'd written for The New Yoker called AI versus m
- 5:46Speaker 1: D which laid out all of the kind of benefits
- 5:48Speaker 1: and potential applications of using AI in medicine, including some
- 5:52Speaker 1: of the downside such as the black box problem of
- 5:54Speaker 1: AI that you mentioned, not knowing why an algorithm has
- 5:57Speaker 1: made a recommendation, and also another problem, which is that
- 6:00Speaker 1: if we really too much on technology, it can erode
- 6:02Speaker 1: human skills. There is a fear that AI could move
- 6:06Speaker 1: us into a very black and white way of thinking.
- 6:08Speaker 1: The computer says you have cancer, or the computer says
- 6:11Speaker 1: you to have your liver removed, said Arthur, who is
- 6:14Speaker 1: one of the world's foremost oncologists and a Pullet Surprise
- 6:16Speaker 1: winning author. Provided a different perspective. There is something very
- 6:21Speaker 1: fundamental about the human brain, a scientist's brain, a doctor's brain,
- 6:25Speaker 1: and artist's brain that asks questions in a fundamentally different manner,
- 6:29Speaker 1: the why question. Why did this happen in this person
- 6:32Speaker 1: in this time? Why does the melanoma appear in the
- 6:35Speaker 1: first calase? What is the molecular basis of that appearance.
- 6:38Speaker 1: The most interesting mysteries of medicine remain mysteries that have
- 6:41Speaker 1: to do with the why. Once we give up some
- 6:45Speaker 1: of the diagnostic pattern recognition material to machines, it will
- 6:49Speaker 1: be time to play. It will be the time to
- 6:50Speaker 1: play in the arena of human therapeutics, human biology, the
- 6:55Speaker 1: complexity of the human interaction, the art of medicine. My
- 6:58Speaker 1: hope is that medicine, in being more playful, will become
- 7:01Speaker 1: more compassionate, more able to take into account individuals and
- 7:06Speaker 1: their individual destinies rather than bucketing people in big categories.
- 7:11Speaker 1: It means having more time to spend with humans. You know,
- 7:15Speaker 1: we are so constrained by time that even compassion gets
- 7:19Speaker 1: three minutes, We won't become more robotic, will become less
- 7:23Speaker 1: robotic as the robots enter our own So Dartha's point
- 7:28Speaker 1: is that these tools could make doctors more efficient so
- 7:32Speaker 1: that they can provide better care. It sort of takes
- 7:34Speaker 1: the grunt work out of medicine and puts the patient
- 7:39Speaker 1: care work back in the doctor's hands. This idea that
- 7:43Speaker 1: technology can actually allow us to be more human, make
- 7:46Speaker 1: us more empathetic, is fascinating, and it also raises the
- 7:50Speaker 1: questions about new types of skills that may need to
- 7:52Speaker 1: be developed in an age of AI. Yeah, and Regina
- 7:55Speaker 1: Barslay from m I T spoke a lot about this.
- 7:58Speaker 1: How doctors have to now equip themselves with new ways
- 8:02Speaker 1: of translating data to patients, we still do not communicate
- 8:06Speaker 1: it to the patient because I think now there is
- 8:10Speaker 1: a walk to be done, not on computer science or AIPAD,
- 8:13Speaker 1: but really on the clinical side. What is the best
- 8:16Speaker 1: way to communicate it to the patient and what is
- 8:20Speaker 1: um you know, the past that you're going to give them.
- 8:23Speaker 1: It is not just enough to say you know you
- 8:25Speaker 1: are high risk. You need to propose some suggestion and solution.
- 8:31Speaker 1: So currently the clinical stuff is thinking and looking at
- 8:35Speaker 1: the ways of effective you know, clinical engagement with a patient.
- 8:40Speaker 1: You know, I speaking of data. You all know Harri
- 8:43Speaker 1: was another person who made you and me think about
- 8:45Speaker 1: humans as reducible to data. I think he's mostly known
- 8:49Speaker 1: as a historian and for his book Sapiens, but you
- 8:52Speaker 1: spoke with him about the data we produce as humans
- 8:55Speaker 1: and how that influences our relationship with technology. That's right,
- 8:58Speaker 1: which is the topic of his book Daus. And he
- 9:00Speaker 1: has this phrase data is m to describe how we've
- 9:03Speaker 1: kind of come to worship the data we create and
- 9:07Speaker 1: our own technological creations. So what happens when based on
- 9:12Speaker 1: all of our past behavior, AI starts to know us
- 9:15Speaker 1: better than we know ourselves. Here's a clip from You
- 9:18Speaker 1: are talking about exactly that. When we talk about AI,
- 9:22Speaker 1: we tend to greatly exaggerate the potential abilities, but at
- 9:26Speaker 1: the same time we also tend to exaggerate the abilities
- 9:30Speaker 1: of humans. People say that AI is not going to
- 9:35Speaker 1: take over our lives because it's very imperfect and it
- 9:39Speaker 1: won't be able to know us perfectly. But what people
- 9:43Speaker 1: forget is that humans often have a very poor understanding
- 9:47Speaker 1: of themselves, of the desires, of their emotions, of their
- 9:52Speaker 1: mental states. For AI to take over your life, it
- 9:57Speaker 1: doesn't need to know you perfectly, just need to know
- 10:01Speaker 1: you better than you know yourself. And that's not very
- 10:04Speaker 1: difficult because we often don't know the most important things
- 10:08Speaker 1: about ourselves. So, but let's say you could turn back
- 10:11Speaker 1: the clock to being fifteen, would you have wanted to
- 10:14Speaker 1: live in a world where there was sufficiently good sensors
- 10:17Speaker 1: to monitor your eyes, your eye movement, you're breathing, you know,
- 10:21Speaker 1: while you're going about your daily life, and then to
- 10:23Speaker 1: interpret that and say to you you have all more
- 10:27Speaker 1: likely than not you're gay. That's a very good question
- 10:30Speaker 1: which will become very practical questions in a few years.
- 10:34Speaker 1: And the way that I grew up and developed. It
- 10:38Speaker 1: would have been a very bad idea. I wouldn't like
- 10:42Speaker 1: to receive this kind of insight from form a machine.
- 10:46Speaker 1: I'm not sure how I would have dealt with it
- 10:49Speaker 1: when I was fifteen, you know, in Israel in the
- 10:51Speaker 1: nineteen eighties, and maybe partly it was you know, a
- 10:54Speaker 1: defense mechanism. In the future too. It it depends where
- 10:57Speaker 1: you live. Brunei has instituted the death penalty for gay people,
- 11:02Speaker 1: at least for people engaged in homosexual sex. So if
- 11:05Speaker 1: I'm a teenager in Brunei, I don't want to be
- 11:08Speaker 1: told by the computer that I'm gay, because the computer
- 11:11Speaker 1: will then be able to tell that to the police
- 11:14Speaker 1: and to the authorities as well. So the apps we use,
- 11:17Speaker 1: the product we buy, the number of steps we take,
- 11:20Speaker 1: the delivery I ordered last night, that will becomes data,
- 11:24Speaker 1: and that data can feed into neural networks to create
- 11:27Speaker 1: statistical models of us and what we might do next,
- 11:30Speaker 1: sometimes in order to diagnose a medical condition, and other
- 11:33Speaker 1: times to sell us a product. Here's V again, looking
- 11:37Speaker 1: to the future, say ten twenty years. The danger is
- 11:41Speaker 1: if I still don't know that I'm gay, but the
- 11:43Speaker 1: government and Coca Cola and and Amazon and Google. They
- 11:47Speaker 1: already know it. I'm at a huge disadvantage. So it
- 11:51Speaker 1: could be something as frightening as the secret police coming
- 11:55Speaker 1: and taking me to a concentration camp. But it could
- 11:58Speaker 1: also be something Coca Cola knowing that I'm gay, they
- 12:03Speaker 1: want to sell me a new drink, and they choose
- 12:06Speaker 1: the advertisement with the shirtless guy and not the advertisement
- 12:10Speaker 1: with the girl in the bikini. And next day morning
- 12:13Speaker 1: I go and I buy this soft drink and I
- 12:15Speaker 1: don't even know why, and they have this huge advantage
- 12:19Speaker 1: over me and can manipulate me in all kinds of ways. Well,
- 12:23Speaker 1: as you've all brought up soda, I was not allowed
- 12:26Speaker 1: to drink soda as a child. My parents tricked me
- 12:30Speaker 1: into thinking that Seltzer was soda. I later found out
- 12:33Speaker 1: that soda is soda and Seltzer is water. And somehow
- 12:38Speaker 1: the Seltzer of it all is the perfect segue because
- 12:41Speaker 1: AI is not only being used to sell a product,
- 12:45Speaker 1: it's also being used to create products like Seltzer in
- 12:49Speaker 1: the R and D research and development phase. And for
- 12:52Speaker 1: this bombs episode, Julian remember Juliana, a lovely producer, And
- 12:57Speaker 1: I went to meet the company behind the gastrograph app,
- 13:00Speaker 1: which is using consumer preference data to make predictions about
- 13:04Speaker 1: new flavors. People might like, I think you hate Nettle,
- 13:08Speaker 1: think against Analytical Flavor Systems or a f S is
- 13:15Speaker 1: tucked away down a side street in Chinatown, up in
- 13:19Speaker 1: a third story walk up. It does actually and when
- 13:23Speaker 1: we were there, the office was still waiting for furniture.
- 13:25Speaker 1: You know, it had this We're going to disrupt the
- 13:27Speaker 1: industry vibe. We moved in like a month or two ago.
- 13:31Speaker 1: But it turns out you can't just buy office furniture.
- 13:36Speaker 1: We met the founder of the company, Jason Cohen, and
- 13:38Speaker 1: we believe that in the future, in order to be competitive,
- 13:41Speaker 1: you have to be targeted. That that there won't be
- 13:43Speaker 1: any billion dollar brands right in ten years if you
- 13:47Speaker 1: don't have an m b A. Here's what Jason's talking about.
- 13:50Speaker 1: Think about the coffee you drank this morning, is it
- 13:52Speaker 1: Third Wave? Did you get it from Starbucks or an
- 13:55Speaker 1: indie roaster. Food and beverage companies are moving more and
- 13:59Speaker 1: more towards knee markets. The problem is that they have
- 14:02Speaker 1: very old school ways of developing new products. But a
- 14:05Speaker 1: FS is offering another way to reach those customers, and
- 14:08Speaker 1: that's by allowing companies to formulate more specific products using AI.
- 14:13Speaker 1: Here's Jason again. Usually the way that things are done
- 14:16Speaker 1: today is you get some conceptual brief. You might say,
- 14:20Speaker 1: we want to develop a new fruit flavored beverage for
- 14:24Speaker 1: Japanese millennial women. Right, you would look at what other
- 14:27Speaker 1: fruit flavored beverages are out there. You'd look at your
- 14:29Speaker 1: own product lineup and say, well, we already have a
- 14:31Speaker 1: lemon flavor, and we're going to send out these briefs,
- 14:35Speaker 1: and we're gonna send these out to these flavor companies.
- 14:36Speaker 1: We're gonna see what other fruits we can get, and
- 14:38Speaker 1: you're gonna wind up with very mainstream things. You're gonna
- 14:41Speaker 1: wind up with peach and mango and strawberry and grapefruit, right,
- 14:45Speaker 1: and maybe you'll wind up with something interesting like low
- 14:48Speaker 1: quad or uzo or dragon fruit. Right. And then you're
- 14:53Speaker 1: gonna have your own consumer tasting panel internally hopefully uh
- 14:57Speaker 1: and you're gonna have them taste it, and they're gonna
- 14:59Speaker 1: have to like some of those more than your current
- 15:02Speaker 1: offering or more than a competitors. After you've done all
- 15:05Speaker 1: of this work, which sometimes costs in the tens of
- 15:07Speaker 1: thousands of dollars in order to have the samples developed.
- 15:09Speaker 1: How the samples sent to you recruit the consumers, but
- 15:11Speaker 1: the product in front of the consumers, that data is
- 15:14Speaker 1: only ever usable once. Right. All you get from that
- 15:17Speaker 1: as a binary yes no, sixty percent of the population
- 15:20Speaker 1: liked more than the competitors, and so what we're doing
- 15:23Speaker 1: is entirely different. Jason's team wants to take product development
- 15:28Speaker 1: out of the yes no binary. Instead of just saying
- 15:31Speaker 1: coke or pepsi, they can calculate which parts of each
- 15:35Speaker 1: soft drink people liked and disliked, and then a f
- 15:38Speaker 1: s can make entirely new flavors based on what a
- 15:41Speaker 1: person kind of liked about pepsi and kind of liked
- 15:44Speaker 1: about coke. And finally, they can transfer those preferences to
- 15:48Speaker 1: entirely different demographics. So what might someone in Mexico want
- 15:52Speaker 1: to taste in their cola compared to what a Japanese
- 15:55Speaker 1: millennial woman might want to taste in her cola? This
- 15:58Speaker 1: is what Jason believes is truly disruptive, being able to
- 16:02Speaker 1: say to a brand, if you want to launch in Mexico,
- 16:04Speaker 1: you should tweak your flavors in this way because we're
- 16:07Speaker 1: actually able to collect this data, develop a data set,
- 16:10Speaker 1: and make predictions. So we could take data from say, white,
- 16:12Speaker 1: twenty to three year old college educated males, and use
- 16:15Speaker 1: that to predict what every other population in the United
- 16:17Speaker 1: States is going to perceive in that product. And so
- 16:19Speaker 1: we're taking an industry that has never seen predictions of
- 16:21Speaker 1: any kind before and finally being able to actually predict things,
- 16:25Speaker 1: predict who's gonna like it, how much they're gonna like
- 16:27Speaker 1: it right, and what we can do to optimize it
- 16:29Speaker 1: so that they like it more. We can actually create
- 16:31Speaker 1: products that no one would have ever thought of, and
- 16:33Speaker 1: no one ever would have thought that a segment of
- 16:35Speaker 1: the population would have liked. And this is something that
- 16:37Speaker 1: we now do with the companies that we work with.
- 16:39Speaker 1: We did talk quite a bit about developing a pine
- 16:41Speaker 1: flavored beverage in Japan. When we first showed these results
- 16:45Speaker 1: to a company there, they said, natzu desca, do you
- 16:47Speaker 1: mean pineapple? Because it was just so out of the yeah,
- 16:50Speaker 1: out of the ordinary. The way Jason and his team
- 16:54Speaker 1: are able to get such nuanced data is with an
- 16:56Speaker 1: app they developed called Gastrograph. Gastrograph looks a lot like
- 17:00Speaker 1: the flavor wheels they use in coffee and wine tasting
- 17:02Speaker 1: to help people map out what they taste when they
- 17:04Speaker 1: try a new product. We think of every flavor, romance,
- 17:07Speaker 1: texture as a signature. You can have the five basic
- 17:10Speaker 1: flavors bitter, sweet, salty, sour mommy, and then underneath that
- 17:13Speaker 1: you're gonna have categorical flavors like fruity earth, the herbaceous nuts,
- 17:16Speaker 1: and seedge roasted. I mentioned that he's a professional taster, right,
- 17:18Speaker 1: and then underneath that you can have subcategorical like citrus,
- 17:22Speaker 1: or specific like lemon are very specific like Meyer lemon
- 17:24Speaker 1: are very very specific like Meyer lemons as right, So
- 17:27Speaker 1: all of those signatures exist in some infinite dimensional space
- 17:32Speaker 1: flavor space. So, car, since you're such a Seltzer fiend,
- 17:36Speaker 1: we demo gaster graph and got a feel for the
- 17:38Speaker 1: app by doing a Seltzer tasting. Yeah, we tasted five
- 17:42Speaker 1: Seltzers that are already on the market that you'd buy,
- 17:44Speaker 1: like seven eleven, and we rated different components of them.
- 17:48Speaker 1: So if I tasted fruit, I'd rate that from zero
- 17:50Speaker 1: to five, and I could add adjectives like strawberry or mango.
- 17:54Speaker 1: You know, I'm not into like Seltzer two point oh
- 17:56Speaker 1: with like kumquat flavored sparkling seltzer. You know, I plane vintage,
- 18:02Speaker 1: but a f s is Resident chemist Ryan On agreed
- 18:05Speaker 1: to formulate a seltzer based on just our extremely small
- 18:09Speaker 1: data set. Um, you'd probably be a pretty fast process.
- 18:12Speaker 1: So we have tons of seltzer data. What we would
- 18:15Speaker 1: want to do is um run through a couple of
- 18:19Speaker 1: different flavors to get an idea of the types of
- 18:20Speaker 1: things that you like, build a model specifically around that,
- 18:24Speaker 1: run an optimization, predict a new flavor that you've never
- 18:27Speaker 1: had before, and then have you tried again. So after
- 18:31Speaker 1: we did all of that, we went home. We sat
- 18:34Speaker 1: in our hands, and then we went back to the office,
- 18:38Speaker 1: which actually a little bit more furniture when we got
- 18:41Speaker 1: back there to see if they actually could create a
- 18:45Speaker 1: seltzer that both Julian and I would like. So it
- 18:48Speaker 1: was a blind tasting. Here we go, all right, drink
- 18:51Speaker 1: one pair. I don't know why I think such a
- 18:56Speaker 1: thing against pears. The number two berries, that's delish. Let's
- 19:02Speaker 1: see what we got here. I wouldn't know huckleberry if
- 19:05Speaker 1: it hit me in the face, honestly, but whatever, Number
- 19:10Speaker 1: three honestly, curry. Do you not taste curry like there's spice?
- 19:18Speaker 1: It's so interesting? One? Oh, I love it like hops
- 19:24Speaker 1: and I'm si care just taste like grass? What are you?
- 19:32Speaker 1: We're done? So the first thing you guys should tell
- 19:35Speaker 1: us is which is your favorite? For me? Is number three?
- 19:40Speaker 1: Just because I like the complexity of flavor. Three was
- 19:42Speaker 1: a beverage too. I'm used to and I've probably had before.
- 19:46Speaker 1: Maybe three. I really enjoyed one I did not like.
- 19:52Speaker 1: I don't like those flavors. We have to reveal. We're
- 19:55Speaker 1: almost at the reveal can to clarify. Your job as
- 19:58Speaker 1: a company is to predict future products that people will
- 20:03Speaker 1: enjoy and come back to. Yes, So in that regard,
- 20:07Speaker 1: three was a winner. All right? Oh my god, tes us.
- 20:13Speaker 1: So we had to do one product that was going
- 20:15Speaker 1: to be optimal for both of you, and we got it.
- 20:18Speaker 1: It was number three. They nailed it. You could say
- 20:24Speaker 1: that we got flavor hacked up. Here. This this blue graph,
- 20:27Speaker 1: this is saying that there's a you know, a seventy
- 20:29Speaker 1: percent chance that you would give this a six? Did
- 20:33Speaker 1: I give it a six? I think I did? You? Did?
- 20:36Speaker 1: You both gave it a six? Um? So we were
- 20:38Speaker 1: pretty confident on this, but we didn't have that level
- 20:40Speaker 1: of confidence that we saw this. What I tasted is
- 20:43Speaker 1: not something I ever predicted I would have liked, but
- 20:47Speaker 1: it's absolutely something I will continue to think about. It
- 20:50Speaker 1: was such a unique flavor and it's actually something I
- 20:53Speaker 1: would buy. It's just that I had never tasted something
- 20:55Speaker 1: like it before. So when I first tasted it, I
- 20:57Speaker 1: was like, this is strange. In the case scenario, gastrographs
- 21:01Speaker 1: AI can help companies create foods that satisfying more specific
- 21:05Speaker 1: tastes and even bring people more joy, and that's good
- 21:08Speaker 1: for business. Instead of making huge bets and trying to
- 21:11Speaker 1: market a product to an entire country, a f S
- 21:14Speaker 1: has created a way to make more specialized bets and
- 21:17Speaker 1: help companies tap into those niches. And this isn't about
- 21:20Speaker 1: AI reducing our experiences to data. AI is being used
- 21:24Speaker 1: to change how we experience the world. More sleepwalkers after
- 21:29Speaker 1: the break So, Karen, I would have been quite nervous
- 21:43Speaker 1: to stand in the shoes of Jason and Ryan and
- 21:45Speaker 1: the gastrograph team because I know that you're such a
- 21:48Speaker 1: connoisseur of Seltzer. How do they do? They did really
- 21:54Speaker 1: well actually, and I think it's important to mention that
- 21:57Speaker 1: they weren't trying to create something they thought I would
- 22:00Speaker 1: already like, like I love cranberry, right. They were trying
- 22:04Speaker 1: to create something that I hadn't tasted before and also liked.
- 22:08Speaker 1: So it was really difficult, and I think it also
- 22:11Speaker 1: shows that there's a bit of reversal in the way
- 22:14Speaker 1: that we do things. Companies have always used market research
- 22:17Speaker 1: to predict consumer preference, but it's often based on things
- 22:20Speaker 1: like focus groups or survey research. What we have now
- 22:25Speaker 1: is massive amounts of data being funneled through an algorithm
- 22:28Speaker 1: to deliver the perfect product for a very specific type
- 22:32Speaker 1: of person exactly. That's the or yeah, age, demographic, socio
- 22:37Speaker 1: economic race. They can target it to all these particular categories,
- 22:42Speaker 1: and I think this is cool and also a little unsettling.
- 22:46Speaker 1: I think as humans, we like to be in control,
- 22:48Speaker 1: you know. I like to think that my preferences are
- 22:51Speaker 1: just that, my own preferences, and this sort of up
- 22:55Speaker 1: ends that notion. You know, using pre existing data, I
- 22:58Speaker 1: can kind of be read they're reading me, and that
- 23:02Speaker 1: makes me feel a little less special. I do think
- 23:05Speaker 1: it's cool that companies are trying to service niche markets,
- 23:08Speaker 1: and I think that's a trend I would definitely get
- 23:10Speaker 1: on board with as far as AI being used to
- 23:13Speaker 1: make predictions. And the reason this gastrograph pieces interesting is
- 23:16Speaker 1: because it's a perfect demonstration that AI is not some
- 23:20Speaker 1: thing which is going to happen in the future. It's
- 23:23Speaker 1: here with us today. We can literally taste it already
- 23:34Speaker 1: AIS in our lives. It's interpreting our data, is analyzing
- 23:37Speaker 1: our preferences, is predicting our behaviors. But we're just starting
- 23:41Speaker 1: to respond to what that means culturally. And so there's
- 23:44Speaker 1: a lot of new technologies and new issues that were
- 23:46Speaker 1: very excited to get our hands dirty with on season two. Absolutely,
- 23:50Speaker 1: and one of the important issues we're going to explore
- 23:53Speaker 1: is bias and technology. It's easy to think that algorithms
- 23:56Speaker 1: are neutral, but the reality is that technology is built
- 23:59Speaker 1: by someone, and that person's bias can be built into
- 24:02Speaker 1: a system. This Princeton professor named Ruha Benjamin has introduced
- 24:07Speaker 1: a concept directly related to this, called the new Gym code,
- 24:11Speaker 1: which asks us to consider the inequities encoded in algorithms. Well,
- 24:15Speaker 1: it's the algorithms and also the data they learned from right.
- 24:18Speaker 1: I mean AI harnesses the power of processing huge amounts
- 24:22Speaker 1: of data about things that have happened in the past
- 24:25Speaker 1: in order to predict a future, and so we have
- 24:27Speaker 1: to be very careful about what that data contains, or
- 24:30Speaker 1: we might not like the future. It's bits out. I
- 24:32Speaker 1: think it's a particularly relevant in the air of medicine.
- 24:35Speaker 1: We see huge, huge promise about honesting AI for better
- 24:38Speaker 1: medical outcomes, but we also need to be very careful
- 24:41Speaker 1: about how the data is being used and who has
- 24:43Speaker 1: access to it, and how can you prevent your data
- 24:46Speaker 1: from being used against you. Well, we have to think
- 24:49Speaker 1: about the potential for data to harm but also to
- 24:53Speaker 1: provide comfort and to drive innovation, sometimes extraordinary, very unexpected innovation.
- 24:59Speaker 1: There are two stories I can't wait to dive into
- 25:00Speaker 1: in season two. One is about a doctor using AI
- 25:03Speaker 1: to record and optimize conversational strategies with very sick patients.
- 25:09Speaker 1: What should they say, how and when? Another is about
- 25:13Speaker 1: using natural language processing to enable immersive conversations with holograms
- 25:18Speaker 1: of people from history, everyone from astronauts to Holocaust survivors.
- 25:22Speaker 1: In other words, using technology to bring history into the
- 25:25Speaker 1: present and ensuring we never forget our past it's wild,
- 25:31Speaker 1: so we're obviously looking forward to seeing you in the
- 25:33Speaker 1: next season. We have a lot of amazing stories lined up.
- 25:36Speaker 1: We'd love to hear from you guys about stories that
- 25:38Speaker 1: you want to hear or subjects that you want to
- 25:40Speaker 1: hear about, So tweet us at Sleepwalker's Pod on Twitter
- 25:44Speaker 1: obviously and on Instagram. We are Sleepwalkers Podcast. Yeah, thank
- 25:49Speaker 1: you so much. We love your feedback and we're really
- 25:51Speaker 1: looking forward to seeing you for season two very soon.
- 26:06Speaker 1: Sleepwalkers is a production of I Heart Radio and Unusual Productions.
- 26:10Speaker 1: For the latest AI news, live interviews, and behind the
- 26:13Speaker 1: scenes footage, find us on Instagram, at Sleepwalker's Podcast or
- 26:17Speaker 1: at Sleepwalker's podcast dot com. Sleepwalkers is hosted by me
- 26:22Speaker 1: Ozveloshin and co hosted by me Kara Price. Were produced
- 26:25Speaker 1: by Julian Weller with help from Jacopo Penzo and Taylor Chcogin,
- 26:29Speaker 1: mixing by Tristan McNeil and Julian Weller. Our story editor
- 26:32Speaker 1: is Matthew Riddle. Sleepwalkers is executive produced by me Ozveloshin
- 26:37Speaker 1: and Mangesh hat Together. For more podcasts from my Heart Radio,
- 26:41Speaker 1: visit the I Heart Radio app Apple podcasts, or wherever
- 26:44Speaker 1: you listen to your favorite shows.