[Music] [Applause] [Music] ai could help us in so many ways if we get it right it could help us solve climate change um it could help us address um inequities in the criminal justice system it could help it could help with us educate many many more people with the standards you know that students at a school like like uh would get it could really really change things for the better [Music] [Music] [Music] hi [Music] a [Music] away mahalo noel for that lovely olay and welcoming us here to the hawaii uh manoa valley and mahalo for all of you that have joined us tonight i'm denise conan i'm an economist and i'm the dean of the college of social sciences here at university of hawaii at manoa and this is just a very special occasion because this is the first in-person event that is part of the uh better tomorrow speaker series in the past two years and we've all been through challenges this past time and it is quite special that we've gathered here in person um i know that you're all here and and feeling that as well and i do want to also acknowledge that in addition to people who have come here tonight in person we have many many that are watching us online and we welcome you all thank you so much for joining and tonight really is about togetherness in a sense and new ways of coming together in this digital age this way that technology is bridging divides is even through the lockdown just has been tremendous and we've all adapted in incredible ways and even as we're marveling through new apps that help us through these times we know that it's really changing us as a society in some ways it's bringing us closer together in some ways there are divides and digital divides that may mislead us may make it difficult and lots of innovations that are happening in our lives in response to times like this and so we're so fortunate that we are going to be able to as a university like ours take on some of these issues and discuss these issues here tonight as a new technology shaped the world how can we best prepare our students for their lives in this new era that we are all still seeking to discover um so in events like this the um it's it's just really wonderful to see that the better tomorrow speaker series is able to step in and provide insights for us this is a joint venture of the university of hawaii manoa the hawaii community foundation kamehameha schools and it's also something that many in the university are supporting my college college of social sciences you hero the scholar strategy network and the william s richardson school of law all participated to make tonight's events possible and so i really want to thank you all for coming out i want to especially thank our student staff who have worked really very very hard if you could stand up those that have worked hard to make this event happen in the back thank you so much and the bookstore too for those of you that were able to get a book here tonight um so the better to tomorrow's speaker series aspires to not just spotlight good ideas but to help move us and nudge us forward in new areas so we try to include um as we're planning not just our faculty and students who are here tonight but our public officials and we're so grateful for that i want to thank i want to thank michael bruno provost and not acknowledge that he's joined us here this evening and at this time i'm very very honored and pleased to introduce lieutenant governor josh green to the stage who will be introducing our speaker we're very honored because you know actually um josh you introduced our last in-person event to it remember back in time um two years ago and we're so glad that you can be with us here this evening to um to introduce the speaker here tonight we owe a heartfelt gratitude to you for all you've done for our state in leading us through these very very challenging times and we welcome you i welcome you to this stage thank you so much dr greene [Applause] so we've begun to kiss again though we are highly boosted as a population here on the stage so thank you guys uh for welcoming me so warmly uh thank you for surviving covid and for wrestling with what uh was obviously the toughest two years of our lives uh here in hawaii uh i feel a great honor to be able to be with you and i hope it's okay um uh denise if i ask for just um a brief before we introduce our esteemed guest who is really extraordinary uh kevin roos if we could just ask for a moment of silence to honor paul farmer who was exceptional as a human being let's give him a round of applause [Applause] we'll take a moment of silence for his passing and he would doubtless scold us for uh being sad about anything as uh life goes on and he promoted life for millions and we promote wisdom tonight and thoughtful leadership for the democratic process of people guiding this world this technology world but i'm going to get into kevin now who is exceptional already so yes it was two years ago when we started the better tomorrow series uh with our in our last np uh in person speaker harvard psychologist robert waldinger and it was this auditorium so you're back here again continuing that legacy uh tonight's a different talk it's about the issues of momentous importance that we see shaping now our lives we're talking about how we look at our screens our relationships with a different world and how apps bring us knowledge and how they impact what we do and how we're connected it's a powerful technological question that we'll be able to hear about tonight from an expert it affects our relationships and our relationships have been perhaps clouded by a mask and they've been behind closed doors and so i'm happy to see the kind of this awareness come out tonight now let's talk a little bit about kevin ruse kevin is a quaker who was raised by swarthmore parents which is very important to me because i went to swarthmore no one knows what swarthmore is because tiny little college in the middle of nowhere outside of pennsylvania but it's not surprising to me uh to feel kind of the gentle nature that he has coming from those beginnings i think you're going to see that reflected in his conversation tonight he seems like a just a genuinely good person he's also here he doesn't look like he's under duress but he is expecting a baby boy he and his wife in the next like 12 minutes so let's give him a round of applause and pray they don't go into labor tonight back in california but he will offer us thought about the practical recommendations on every level of action in this space he will talk to us about educational institutions like ours and how how we actually can benefit from broad-based liberal education in this climate in this world he went to brown which i think is inferior to swarthmore in every way mostly because my wife went there and my my very famous uncle went there and so we like to tease those brown folks but um when i see them come out and and talk to us about these emerging technologies and liberal art things it actually i tease of course it's so wonderful to see the value continue to place to be placed on higher education in the social science and the humanities and it brings us together we saw our music director here tonight and we see deans come to talk so it becomes a special process when we get someone like kevin rouse here so all the recommendations he offers us tonight i assume will often be hopeful rather than dystopian but i don't know what he'll come up with we've had plenty of dystopia so it has been tough but i really am so pleased that we have someone of his prominence he grew up in a small town in ohio wrote his first book as a sophomore in college which is extremely impressive produced a documentary series hosts a podcast which i'm not still not sure exactly i understand how to use podcasts but people always tell me to do them and has written three books already he's lucid he's witty he's just thoughtful he writes for the new york times which we all love and we're really just so pleased to have you here kevin so let's give him a round of applause and celebrate a scholar here with us in hawaii [Applause] aloha thank you for the wonderful uh introduction lieutenant governor um so happy to be here this is a very exciting day for me for a few reasons one of which as uh as lieutenant governor greene noted um my wife is uh somewhere between uh 37 and 38 weeks pregnant she uh is in the danger zone so uh fair warning uh if my watch buzzes mid-sentence and i get the right text i am out of here and you guys are gonna have to entertain yourselves tonight um very exciting it's our first kid and um my wife was bummed because she wanted to come with me here to hawaii one of her favorite places to visit so if you all could do me a favor and just like go on my instagram and just comment about how awful the weather was and oh you can't believe they it rained the whole time and her husband definitely did not drink pina coladas on the beach instead of helping her set up the nursery i'd really appreciate it um the second exciting thing about today is that it's in person we are here in a building with each other uh it is amazing i've done so many events and all through zoom and this is way better already um it's uh i i don't have to like worry about whether i'm on mute before i start talking um if you all pick your noses you can't turn off your cameras i can see you so we've got that uh going for us and the third reason that i'm excited today is that i get to talk to you all um this such a a great group of people um in my day job i i write for the new york times and that means effectively writing for everyone in the english-speaking world um young people old people rich people poor people people who are tech experts and people who got lost in the text section on their way to wordle i talk to them all but you are students and educators and leaders you get it i don't have to tell you that technology or the internet is important or that it's changing society or that you should be paying attention to it you're not scared of the future you've just been through a freaking pandemic and so you are just wondering what is going to happen next and how you're going to get through it so i thought we would just talk a little bit about that and i'll i'll share some things that i learned in the course of of writing this book um and then we can have some questions and conversations now i should say that my my book it's got you know it's got this title it makes it it's got nine rules in it makes it sound like a self-help book and i guess in some ways it kind of is but i am not an advice writer like self-help is not my thing i am a tech columnist i write about iphones and social media apps and cryptocurrencies named after dogs but i wrote this book honestly because of something that happened to me just after i graduated from college so this was about a decade ago i was in my first job in journalism and i was a financial reporter i was covering big banks and and one of the tasks that i had every quarter when the companies would release their earnings reports i would write them up so you know uh morgan stanley would say we made 1.2 billion dollars in profit this quarter and i would go on the release and i would find the right numbers and i would put him into a story and i'd type it up real quick and i'd hit publish and that was like part of my job and it was not the most fun part but it was an interesting experience and i learned a lot and then one day i got an email from this company called automated insights it's automated insights what's that well it's a software company in north carolina and they had just came come up with a new ai reporting app so i perked up and i said well what can this reporting app do and they said well it can write corporate earnings stories it can take data from press releases and plug it into a template and publish it with no human intervention required and it could do all this in milliseconds and they bragged in this email that their software had been used to produce 300 million stories in a single year which is slightly more than i write so i had a problem obviously which is that a piece of software was doing my job better and faster and more accurately than me it was like you know seeing someone who enrolls in your school who is like your exact body double but also like is a rhodes scholar and the captain of the football team and a world-class musician and funnier and more personable than you and he never sleeps of course that's not what i told myself i told myself well i'm i'm not going to be automated i have special skills i'm i'm irreplaceable surely there are things that i can do that the robots can't this is an extremely common reaction in fact if you survey people in the u.s and you ask them the question will robots displace large numbers of american workers in the next 10 years about 75 percent of them say yes they will and then if you ask them will robots displace you in the next 10 years only about 25 percent of people say yes so we have this idea that ai and machine learning and you know robots can do enormously sophisticated things and are going to displace millions of people create mass unemployment but not for me i'm special and when i started reporting this book i i went back and i looked throughout history and i saw this same reaction happening for hundreds of years when the first uh factory equipment arrived during the industrial revolution there were people who said oh they will never have the je ne sais sequoia that these human textile producers have and we know what happened there when the first computers were used to translate languages all the human translators said well they will never be able to capture the nuances of translation now you know we still have human interpreters but google translates getting pretty darn good my favorite example of this was in 1984 the jfk airport in new york introduced the first automated ticket machine where you could go up and it's like you press a button it's like a vending machine for airline tickets and they wrote a story about this they sent the reporter out and they sent him to interview a bunch of travel agents and one of the travel agents gave this quote that i just love he said well this will never catch on what if you just press the wrong button you end up going to a totally different city he could not imagine a future in which people entrusted computers to purchase airline tickets and yes we do have travel agents still but um when was the last time you met one so the problem that we have today this this this displacement cycle has been happening for hundreds of years but the problem today is that it is happening so quickly in my book i have tons of examples of machine learning apps that are doing incredible things that they couldn't do even 18 months or two years ago there are now machine learning models that can diagnose certain types of cancer on scans much more accurately than even the best human radiologists there are ais that can spot issues in legal documents with much higher accuracy than even the best white shoe lawyers they're even you know apps that are creating art and building video game levels and and writing music that is indistinguishable and in some cases better than the analog human counterparts and i could spend the rest of this talk tonight telling you about these ais and what they're capable of doing and how it poses a threat to all kinds of people in their jobs but there's a rule in journalism that says show don't tell so i thought we could do a fun little demo and try to automate the job that i am doing right now so i've been recently i've been playing around with this uh something called gpt-3 has anyone used gpt3 gpt3 is what's called a large language model it's a basically a super powered ai that writes and analyzes text so basically if you can imagine the autocomplete feature on your iphone but then run it with one of the five biggest super computers in the world and feed trillions of examples through it and tell it to be able to to learn how to write text in any number of styles it has 285 000 cpu cores in it so it basically works is you put in some text and you say complete this text and it will write you more paragraphs in any style that you want to you don't have to train it on different authors specifically it is ingested basically the entire internet and it can mimic almost any style it's really good it's so good in fact that open ai the company that developed it didn't release it to the public because they were they were worried it was too powerful and that it would be used to write for example ultra realistic propaganda that would be indistinguishable from propaganda created by humans so gpt3 is not publicly available but i did manage to get a little demo of it there's a product called pseudoright which was developed by an ai engineer in san francisco and who was a frustrated sci-fi science fiction writer and uh he was getting stuck getting writer's block and so he said maybe gpt3 could help so the way it works is you plug in some text and you push this little button called wormhole and it figures out the rest and then you can say i want that to sound more hopeful or i want that to sound uh you know grave or i want it to i want it to sound uh i want it to sound whimsical and it'll reshape the text um to whatever you want so uh this morning i plugged some paragraphs from the speech into gpt3 and i asked it to write me a paragraph about the subject of ai and automation that was suitable for delivery to this audience so i'm just going to read you the output of the machine here automation and ai are the future and we need to prepare for them automation and ai are not inherently evil and we shouldn't fear them and we shouldn't fight them we need to let it enhance our lives and we need to let it create new jobs so we can do some more stimulating and entertaining things because our lives are meant to be more than just work pretty good pretty good i'm sweating a little bit up here and then just to give it a challenge i asked it to rewrite that paragraph in the style of william shakespeare so here's what here's uh gpt-3 shakespeare sounds like oh humanity thou art bound to fall to thy doom thou shalt become slaves of the machines they shall make thee toil endless hours they shall not give thee leave to rest thou shalt toil evermore till thou shalt die whoa and then just for fun i asked it to write me a closing paragraph that sounded hawaiian that's the only uh instructions i gave it i was a little unsure about whether it would be able to do this here's what i came up with we need to bodyboard and surf and watch the blue sea glisten in the shimmering light ohana the ocean shall provide for us pretty good so there you have it and ai can not only do my job right now but it can be shakespeare and it can be a hawaiian poet um i am officially screwed next year you don't even have to invite speakers um and what's fascinating about gpt3 is that it's not just a threat to writers um the open ai the company that developed it last summer released a version called codex and what codex does is essentially the same thing but instead of producing paragraphs of text it produces working code so you can tell it write me a javascript app that will you know display the price of bitcoin in 20 different currencies with a green background and a little thing in the corner that tells me the weather and it will build that on the spot and this was this technology is very new but it's improving very very quickly and a lot of these ais out there in different fields are also improving very quickly to the extent that some of the biggest and most prestigious firms in america think that we are on the precipice of a true crisis mckinsey the consulting firm predicts that by the year 2030 45 million americans will lose their jobs because of automation and contrary to sort of the popular narrative about this it's not just going to be truck drivers and fast food workers it will be doctors bankers consultants musicians journalists in fact there are some studies that have suggested that white-collar jobs are actually easier to automate than many blue-collar jobs so uh you mentioned uh some needing some optimism that was officially the depressing part this is the hopeful part so the question for all of us then becomes in this world of rapidly improving machine intelligence what can we do that can't be replicated by machines what is left for us and figuring out the answer to this question is super important because it's not just about like what we do to stay employed as adults it's for students trying to figure out what skills to develop what subjects to study what internships and jobs to aim for it's important for setting our goals and trying to figure out how we can be not just professionally successful but happy and fulfilled in a world where we're competing with machines every day when i started writing this book i started by calling all of the ai experts i knew people in tech companies and government people in research labs and i asked them what advice would you give to someone like me who is worried about being replaced by a machine and what i thought they would say is that i needed to learn how to code basically i needed to go back to school and major in a stem subject i needed to hustle i needed to optimize my life to become as efficient as possible and hyper productive but instead they kind of surprised me because they said essentially the opposite they told me that they would advise that person me to focus on deeply human skills the kind that can't be automated and then i asked the next follow-up question which was well what are those skills what can't be automated what where do humans have a natural advantage and they gave me three answers the first is what i call surprising work ai and computers in general they really like regularity and rules they're very good at doing repetitive structured tasks but they're they're very bad at what's called zero shot learning they don't like thinking on the fly they don't like encountering new situations when they've never seen these kinds of things before or an environment where the rules are constantly changing like this is why ai is very good at playing chess for example because chess is the same game every time it's a bounded board it can practice over and over and over again millions of times getting a little bit better each time but if you asked an ai to like teach a kindergarten class it would fail miserably because that is a job that is all surprises and chaos there's a famous experiment in computer science where programmers created this algorithm that could identify household objects in photographs so you'd feed it a picture of a living room and it would say that's a coffee cup that's a tv that's a chair and so on and they got really this algorithm got really good it was above 95 accuracy on identifying household objects and then they did something they gave it a twist they introduced a photo of a living room with a baby elephant in it and they asked the computer to analyze it and label the objects and the computer basically had a nervous breakdown it not only couldn't identify the baby elephant but it forgot the labels for most of what else was in the room it could not process the context of a baby elephant appearing in someone's living room it was unable to cope with that surprise machines hate surprises so that work the work that involves surprises and chaos and changing environments that's where we have an advantage another area where we have a natural advantage is what i call social work and this is work that basically instead of making things you're making people feel things you're creating experiences and connections and emotions the value of social work is based on meeting social and emotional needs rather than material needs so obviously this would include jobs like therapists and nurses and teachers but it also includes things that you wouldn't necessarily think of as social jobs one example that i wrote about in my book is my accountant this guy named russ he does my taxes every year actually i'm late to call him this year and he is in a field that is highly prone to automation in fact some studies have suggested that tax accounting is literally the single most automatable job in the world because of turbo tax and everything else and russ is a pretty good accountant but the reason that he's survived is not because he's good at doing taxes it's because he's a former stand-up comedian before he got into accounting he wrote comedy performed comedy he hires other comedians and and you know pays for them to get their their their accounting degrees and he has built a firm of extremely funny accountants people do not believe me when i say that i enjoy doing my taxes but it is true i have more fun itemizing my deductions with russ than i have had at actual comedy shows that i've paid money to see and i asked russ like why did you do this why did you hire a bunch of comedians and train them to be accountants why wouldn't you just hire accountants and he said because we're not selling accounting we're selling the feeling that comes with having someone who is competent and personable and trustworthy help you manage your money we are solving a human need rather than a material need and just by giving that little twist to his outlook he has managed to escape the robot apocalypse the third category of work that is protected or relatively protected is what i call scarce work and scarce work doesn't mean that there are only a few of these jobs in fact i think there are quite a lot but they're jobs that require sort of rare skills or combinations of skills activities that are high stakes or dangerous to automate jobs with a low fault tolerance so for example today uh if you call 9-1-1 to report an emergency a human picks up why is that it's not because we don't have the technology to automate that job we have automated phone trees you could get a system that says this is the police press one to report a burglary press two to report you know something some other crime but we've decided collectively as a society that that job is too important to automate we don't want to fumble through a menu we don't want to press zero to talk to a representative something's going wrong if we're calling 9-1-1 and we have decided that we need someone on the other end of the line who can understand contextualize and respond to our our needs right there and then a less morbid example of a scarce job is uh is my colleague will shorts who is the new york times's in-house cruciverbalist anyone know what a crucial verbalist is he edits the crossword puzzle and a fun fact about will shorts is that he is the only person in the world who has gotten a college degree in enigmatology which is the study of puzzles he's a huge nerd and because he's been doing this for his whole career he has written and edited thousands and thousands of crossword puzzles encyclopedic knowledge of the entire world of crossword puzzles if you took him out in a comment you know a comet took him out i don't know how he would go on as a paper he is literally irreplaceable because he's managed to develop an extremely specific skill and develop it at a high level and so if you put these categories together surprising social and scarce you start to get an idea of like what kinds of people are well equipped to survive this technological transformation they're the kinds of people with empathy and emotional intelligence the kinds who are good at reading a room and meeting people where they are figuring out how someone is feeling and responding appropriately they're people who make experiences rather than things they're people who follow their passions no matter how esoteric or crazy they sound they're people with moral courage people who are cool in a crisis who can handle surprises and people who when presented with a challenge like for example going to school in the middle of a pandemic embrace that challenge and figure out the best way through it for years those skills these deeply human skills that make us stand out in a world full of intelligent ais have gotten a bad rap we called them soft skills i hate that term we taught students that the only people with job prospects the only people who would end up succeeding and having stable careers were coders and engineers but i'm confident that that is wrong that that is almost exactly backwards because i live with those coders i mean i live in the bay area i know tech executives i they're some of my friends and that's not how they're raising their kids their kids go to the waldorf school where they play with blocks in the dirt and they have they have no access to screens at all their kids are being taught that emotional intelligence and empathy are the skills of the future the ones that they will need to succeed in a highly automated economy and i think they may know something that the rest of us don't they know that the world is changing and that the jobs that they trained for aren't necessarily going to be around 10 or 20 years from now they know that even the programmers who have still pretty good job prospects right now will face the same kinds of challenges as people in other fields and that what will be left for us after the gpt-3s of the world have their way with us is the human work so uh i want to share just a few thoughts about that human work before we wrap up tonight and these are things that i came across in the process of reporting this book that i think can help us understand that more of the how the how we transition to an economy and a society that values human work appropriately and the first is that i think we need to radically humanize our educational system you'll often hear people in the humanities talking about how it's a moral responsibility to give young people an access to a liberal arts education which you can agree or disagree with that but forget morals for a second i'm talking about the fact that it is a strategic advantage for young people to learn the skills the deeply human skills that will help them stand out in today's job market and this doesn't necessarily mean that everyone needs to drop out of stem programs and become a philosophy major you can have we need those technical skills too those are super super important but we need to layer on to them a level of human ability and connection that will allow those people to thrive in the future the the coder the sort of stereotype of the coder who sits in a dark room by his or herself pecking out code in isolation that is going away what is needed are the people who can both code and lead a team and collaborate and come up with creative new products and you're starting to see more people in silicon valley acknowledge this that we may have made a mistake in placing so much uh so much importance so much priority on these sort of quote unquote hard skills and letting these quote unquote soft skills kind of erode jack dorsey the founder of twitter recently said that if he could build twitter all over again some of the first people he would have hired would be a behavioral economist and a game theorist because they had a bunch of engineers they built this thing and then the world got its hands on it and did things that they never predicted or expected and he said having a behavioral economist and a game theorist on staff would have helped us anticipate how our service could be misused or what incentives might get people to have healthier conversations on twitter i hear from ceos in tech almost every day who say things like i have all the engineers i need but i can't hire people who can close a sale or write a persuasive memo or win the trust of a client humanizing our educational system doesn't mean eliminating science or technology from our curriculum we need that badly but we also need these other skills and there are some interesting experiments happening with this around the country a lot of medical schools are now teaching classes in medical communication there's a program called oncotalk that just teaches medical students how to talk about cancer not how to diagnose it not how to treat it but just how to talk to people with cancer near the end of their life in a compassionate way that's not a skill that we taught in med schools for a long time but now we're starting to stanford now has an institute for human-centered artificial intelligence where they talk not only about how to build ai but how to use it to improve human lives there are courses that have that are jointly taught by philosophers and computer scientists the university of texas at austin just introduced a course called ethical foundations of computer science which is basically a course that even before you start studying computer science you start with the ethics why you're building what you're building my second thought is that we need to make efficiency a dirty word in our society has anyone ever complimented you on being efficient it kind of stings doesn't it like if i passed out surveys after this talk and asked you all what you thought of my speech and you said that it was efficient i think i would be pretty depressed about that not because i want to waste your time or like spend a bunch of extra time saying things that you don't need to hear but because what you're here for is not efficiency efficiency is a code word that means machine like machines are supposed to be efficient but humans aren't we're supposed to wander and dawdle and have epiphanies while walking through the park we're supposed to be alive and creative we're not supposed to be robots stamping out parts in a factory the most important things in life like climbing mountains and running marathons or having children are not efficient at all and that's why they're meaningful and when we try to make something more efficient like amazon trying to get a package to you in one day instead of two we often mean stripping the humanity out of it treating the humans involved more like robots there's a principle in social psychology called the effort heuristic and the effort heuristic says basically i'm paraphrasing here that the harder we think someone worked on something the more we appreciate it they've run studies for example where they would they take two groups and they give one they give them identical bags of candy and on one group's bag they have a tag on it on the bag that says this candy was personally selected for you by john or whoever the other group has no tag on their on their bag and they have them eat the candy and review the candy and far and away the group with the little tag on their candy reports that it's tastier they like it more they want they feel like someone selected it personally for them and that actually improves the subjective experience of eating the candy so this principle is now everywhere in our economy we see this everywhere we see this in art we see this in entertainment we see this in in influencers and and online culture um the effort heuristic is behind so many of the things that we pay money for whether it's a farm to table restaurant that comes out and explains you know this kale was picked from the rooftop garden or they tell you a story about it because what you're paying for is not actually the kale it's the effort that it took to get that kale onto your plate so i think efficiency is the opposite of the effort heuristic efficiency would say what is quick is good and the effort heuristic would say what somebody worked on hard is good um i got this idea this idea came to me after i was watching a talk by the head of ai research at uh at facebook now meta um who was using the example of a of a dvd player and a ceramic bowl and he said a dvd player is all made by robots uh it has like you know lasers and stuff in it it's a complicated complicated piece of machinery but because it's entirely built by robots it's pretty cheap you can buy one for like 50 bucks but a ceramic bowl if you want a good one that's going to cost you a lot more than 50 bucks and that is a technology that has been around for thousands of years because what you're paying for is the artisan you're paying for the time and the expertise and the labor so i think what i'm saying by we should make efficiency a dirty word in our society is not that everyone should like lollygag or that we should all blow our deadlines and not get any work done but i am saying that i think a good shortcut to figuring out if something is deeply human or not is to consider whether or not it's efficient and i think we also need a radical change in the way that we think about and measure our human efficiency we have today a whole framework around efficiency that you know gross domestic product and economic productivity we we assume that the more gets done in the short time the better off we are but in this new economy that we're moving into this ai economy we need to start measuring our accomplishments in other ways the final thought i want to leave you with tonight and then we'll open it up to questions is something that um a friend of mine who's an ai engineer told me after this book came out we were taking a walk in in oakland where i live and he said you know humans should really get better at distributed intelligence said go on he said well you know the way that ai learns is that there are all these nodes in a network and once one node in the network figures out something it passes it back to all the other nodes so this is why like for example every self-driving car doesn't need to know doesn't need to independently learn what a motorcycle looks like once one of them learns it sends it back to the to the machine learning model and that teaches all the other ones they learn together as a group that's what distributed intelligence looks like but humans don't really do that very well we hoard information we silo it away and you know schools and programs we we don't do an awfully good job of sharing what we learn with others and i think we really need to change that because they are all they're all performing on one team and we are many different teams so i guess my suggestion i want to leave you with as we all inch back out into this irl world into this long period of of isolation is to really share with each other like if you learn something cool tell us to tell it to someone else like build a friendship share a book recommendation mentor someone at work or at school because we are not going to stay ahead of the machines alone we can only do it together thank you so the question if you couldn't hear or you're watching remotely was about gpt 3 this ai writing tool and how you can use it to write your term papers and also what what writers and teachers of writers need to tell them about whether it's still useful to learn how to write i hope it's i sure hope it's still useful to learn how to write because if it's not uh i am i am going to be having a great time as a barista for the rest of my career um i i think what what computers can do and it's very telling that the way this works is that you you feed it an idea and it finishes it computers aren't very good at coming up with ideas so i think there will be a future in which writers and and ais work together and say you know writer comes up with a you know an idea and writes a little bit of it and then there's sort of a handoff at some point um i actually wrote a review for the new york times book review using this tool um a few weeks ago which was a grand experiment and it was you know we disclosed it in the paper that this was written by a robot partially but it was i i would like to think that it wasn't quite as good as me but it was it was passable and so i think there will be certain types of writing that will be automated but there will still be such a need for persuasive communication in every field i mean this is maybe the thing i hear most from executives they ask me they ask me which writers they should hire because they're they're these big executives of billion dollar tech companies and they need people in their companies who can write persuasively and if they could use ai to do it they would be using ai to do it but they need they need the human touch so i think it's still worthwhile uh the next question is also about education what did the pandemic teach us about education i mean you all are much more qualified than i am to answer that question my impression from having you know friends who are educators and and friends with children who have been educated through this pandemic um is that there are a lot of things that just don't really work over a distance um so much of the classroom experience is the moments between lessons the you know the sort of reading the body language of the student who wants to pipe up but is a little bit afraid um you know encouraging collaboration in the classroom these things are you know teachers and educators have been heroes throughout this pandemic at finding ways to adapt um but we know that it's not a perfect substitute i mean all the research shows us that it's not a perfect substitute and that's something and potentially a lot is lost when you go when you go virtual so i think that educators the the great ones are not just transferring knowledge that's not a big part of their job at all actually is like i know something and now i'm going to tell you that something it's a complex emotional relationship and that's really really hard to capture over overzoom um do you think it's possible to humanize education through zoom it's certainly possible to make it more human and i think teachers and educators of all kinds have been really creative about finding ways to adapt to that format but it's hard there are some natural limitations of that model so i don't think it's a i don't think it's going to solve everything next question is i'm a principal at a school i just got hired at google should i take it wow how much are they paying you no um i don't know i if i you know i think it depends on what the job is it depends uh you know what you'd be doing if you feel like you'd be having it it depends what your life is circumstances are like and you know i i never i'm never very good at career advice because my answer is always kind of like it depends but i have i have a huge admiration for educators who should be making the big decisions about technology and how um the big decisions about technology and that includes so much um i think what we're seeing now is some of the consequences of of letting companies make all the decisions about technology i think that you know in past uh technological revolutions what you've seen is a lot more public sector involvement um you know in in the 19 in the early 20th century as in the sort of so-called second industrial revolution the us government took an extremely active role in developing um and electrifying parts of the country they went around to rural towns and turned on the lights and it was just a huge public works project and the internet was really a public sector project um so i think the public sector has been kind of sleeping at the wheel and has not really known how to sort of get its arms around some of these problems and i don't think that's a failure of any individual politician but i do think we we need to start seeing more democratic participation in this process because what we've seen now is essentially these massive you know supranational corporations that have billions of users that kind of act like governments but they're not and they're not accountable like governments are and so i think we need to insert the democratic process in how has reporting on tech changed since the early 2000s with companies like facebook or other platforms catering information has the method of reporting changed well i was uh in high school in the 2000s so i wasn't a tech reporter yet but but i i think it's gotten in some ways much easier in some ways much harder so you know i have access as a reporter to communities and pieces of information and people that i never would have had before you know i can i can look people up it's you know linkedin is a great reporting tool in the old days like if i wanted to contact you know 100 employees of google i would have to like look in the phone book or something and now i can just go on linkedin and say like these hundred people i'm gonna send messages to um so that part of it has gotten easier another place where i really appreciate ai in my daily job is transcription so when i was a reporter when i was a young reporter i had to transcribe all my own interviews and it was the least favorite part of my job because you'd have a two hour interview and then you sit there for six hours typing up the the notes from the tape and now ai just does that instantaneously so that's something that i really appreciate do you think ai could free us from labor do i think ai could free us from labor um it's possible it's that was the sort of john maynard keynes had this uh you know it was like an economist um had this sort of famous essay about how you know in the future we'd all be working you know much much less than we do now i think 20 hours a week and we spend all our free time making art and doing you know philosophizing and spending time with our families and there's a book that came out a few years ago with a funny title called fully automated luxury communism and it's a it's a book about basically this a world in which we are essentially freed from the burden of labor and that we have everything we need because the robots make all of our food and you know give us all of our material goods um i i think a lot of people like to labor as much as we um you know as much as we sort of overwork people and and as much as we should you know all have a more balanced life um i don't know that this has been my sort of contention andrew yang is a friend of mine and and we've argued about this he thinks you know give everyone universal basic income and and uh and i i think there's a lot of merit to that but i also think that people genuinely some some people genuinely like to work they find purpose in it they find a sense of self and community so i i don't know i think it would free us from certain types of labor but i don't think people will ever stop working this question reads i appreciate and agree with your idea of humanizing education and promoting deeply human traits but what do you think about these approaches quickly getting capitalized or otherwise abused example the ethics of computer science course being seen as an annoying requirement or just a cool way to boost your resume it's certainly possible i think educators who are doing these kinds of courses need to be need to be good at it and need to be compelling and need to make it interesting so that it's not just a chore but i think a lot of this is being sort of driven by students i think students want to learn this stuff they they feel kind of gross about the way that technology has kind of captured their attention and you know sort of manipulated their their tastes and their habits so i think a lot of it's being driven by students but i think i think capture of these kinds of programs is certainly um possible so i think i don't know i think educators just need to be vigilant about making sure that it's like not just seen as like the boring thing that you have to do to get your degree but is actually something interesting does ai risk uh perpetuating racism by relying on historical predistinctions to determine outcomes and how do we assure ethical use of ai yes is the answer to the first part of that question um it's it's pretty well documented now a lot of ai systems are just horribly racist um you know they um mislabel um african-american faces as as animals they you know can't tell apart members of various minority groups there's a my colleague cashmere hill wrote an amazing story a year or two ago about a black man who was uh arrested for a crime that he didn't commit because of facial recognition database had mistakenly uh identified him as someone else this kind of thing is happening at scale all over the country it disproportionately affects the poor a number of states are now using ai to do things like determine benefits eligibility who's eligible for food assistance who's eligible for public housing um and when that goes wrong like that has real consequences um for people's lives so these these systems need to be better and they need to the states and and jurisdictions that are implementing them need to really really do their homework before they start turning stuff over to ai this is our final question um what is your optimistic and ideal vision for our future fully automated luxury communism [Laughter] no i mean i i am actually quite optimistic and what what makes me optimistic is honestly like doing things like this talking to students um i i think um there's a great magazine called kernel that i just started getting it's written by these young technologists um and it's sort of like an old school zine if you remember zine's it was like it's like you know it's probably got 400 subscribers or something but i i get it sent me i love it and it's so it every time i pick it up it makes me feel optimistic because the young technologists are really thinking about this stuff and they're really good and i just can't wait till they're running all the tech companies and so i think i am optimistic i think a vision of the future i mean ai could help us in so many ways if we get it right it could help us solve climate change um it could help us address um inequities in the criminal justice system it could help um it could help with us educate many many more people um with the standards you know that students at a school like like uh would get um it could really really change things for the better i mean already i do an annual column called the good tech awards where i like find a you know a handful of things that make me feel optimistic this is like my attempt at uh at keeping myself uh sunny and and optimistic so i find a couple things every year and one of the things i wrote about this year was this project at google um at a subsidiary of theirs called deepmind where they basically solved a protein folding problem using ai protein folding has is this notoriously hard thing that has boggled the top you know cellular biologists for ages and they trained a computer to do it and they solved it and now those solutions are being used to develop hundreds of new drugs the mrna vaccine is a freaking miracle of science and technological progress um so i that's that stuff makes me super optimistic and i i think that's something that i need to do you know a good job of keeping in mind thank you all so much thanks for submitting questions kevin thank you for speaking and for being here thank you all for coming up [Music] you
2022-03-17