The break of call content varification

 

the world is about to change in a fundamental way except the way it's got to change it's not being deployed in a safe and responsible way it's being deployed in a very dangerous way 50% of AI researchers believe there's a 10% for greater chance that humans go extinct from our inability to control ai

the Co founders of the centre for humane technology they were behind the award winning Netflix documentary the social dilemma the social dilemma reached 100 million people in 190 countries in 30 languages and they've also advised for the heads of state global policymakers members of Congress national security leaders in addition to mobilising the millions of us about these issues and some of the dangers that we face with the technology these days so here they are

 

The rubber band effect

this is the first time AI  made me feel something. There was a threshold that we crossed. this was January of last year at that point there were maybe 100 people playing with this like new technology now there are 10 million people having generated over a billion images and trying to explain to reporters what was about to happen and we walked them through how the technology worked and that you would type in some text and it would make an image that had never been seen before and they're not along at the end they be like cool and what was the image database that you got your images from it was just clear that we'd like stretched their mind like a rubber band and then because this was a brand new capability brand new paradigm their minds would snap back and it's not like dumb reporters it's like a thing that we all experience and even in making this presentation so many times relies we have to expand our minds and then we look somewhere else if it snaps back and we just wanted to name that experience because if you're anything like us that will happen to your minds throughout this presentation especially at the end when you go home you believe what did we just see and I think because artificial intelligence is such such an abstract thing and it affects so many things and doesn't have the grounding metaphors like the kinesthetic experience in our lives it's so hard to kind of wrap your head around how transformation this paradigmatic technology what we really want to do is arm you with maybe a more visceral way of experiencing the exponential curves that we're about to be heading and yet you know since 2017 I've been working on thing called their species project using AI to translate animal communication decode non human language so there's a huge part of this stuff that I really love and believe in a couple weeks ago I made a Spanish tutor for myself with chachi BT in like 15 minutes so we're not saying this is great it's better than Duolingo for like 45 minutes So what we're not saying is that there aren't incredible positives that are coming out of this that's not what we're saying that will be our saying his is all the ways that we are now releasing these new large language model a eyes into the public are we doing that responsibly and what we're hearing from people is that we're not doing responsibly with the feeling that I've had personally just to share is it's like it's 1944 and you get a call from Robert Oppenheimer inside this thing called the Manhattan projects you have no idea what that is and he says the world is about to change in a fundamental way except the way it's about to change it's not being deployed in a safe and responsible way it's being deployed in a very dangerous way and will you help from the outside and what I say I've been hammering in more of a metaphor of a large number of people who are concerned about this and some of them might be in this room people who are in the industry and we wanted to figure out what is responsibility look like now why would we say that because this is a stat that took me by surprise 50% of AI researchers believe there's a 10% or greater chance that humans go extinct from our inability to control AI say it one more time half of AI researchers believe there's a 10% or greater chance from humans inability to control yeah that would be like if you're about to get on a plane and 50% of the engineers who make the plane say well if you get on this plane there's a 10% chance that everybody goes down would you get on that plane but we are rapidly on boarding people onto this plane because of from the dynamics that we're going to talk about because sort of three rules of technology that we want to quickly go through with you that relate we're going to talk about this just names the structure of the problem so first when you invent a new technology you uncover a new class of responsibility and it's not always obvious what those responsibilities are so to give two examples we didn't need the right to be forgotten to be written into law until computers could remember us forever it's not at all obvious that cheap storage would mean we'd have to invent new law or we didn't need the right to privacy to be written into law until mass produced cameras came onto the market right brandys had to essentially from scratch invent the right to privacy it's not in the original constitution and of course to Fast forward just a little bit the attention economy we are still in the process of figuring out how to write into law that which the attention economy engagement quantity takes from us so when you went to new technology when cover a new class of responsibility and then two if that technology confers power it will start a race and if you do not coordinate the race will end in tragedy there's no one single player that can stop the race that ends in tragedy and that's really what the social dilemma was about and I would say that social dilemma and social media was actually humanity's first first contact moment between humanity and AI curious if that makes sense to you because it's when you open up tick tock and you scroll your finger you just activated the supercomputer the AI pointed at your brain to calculate and predict with increasing accuracy the perfect thing that will keep you scrolling so we already had we now have every single day in AI which is a very simple technology just calculating what photo what video what cat video what birthday to show you nervous system to keep you scrolling but that fairly simple technology was enough in the first contact with AI to break fake news and breakdown of democracy and no one intended those things to happen right we just had a bunch of engineers who said we're just trying to maximise for engagement it seemed so innocuous and while you're getting better and better recommendations on YouTube that are more and more personalised that YouTube people didn't know that would lead to rabbit holes that sent people into different little micro cults throughout the Internet and So what we want to probably going to talk about what happens in this second contact with AI where we also have a bunch of benefits they were going to get from this technology and there's also erase for for something an easy way to remember that first contact was curation AI second contact creation add generated models all of that and so in this first contact with social media humanity last not now why did we lose how could we have last because we were saying a bunch of things about what social media was right we actually notice we said we're so she is going to give everyone a voice the point here is just like we said there's a paradigmatic response today I what was the paradigm from which we were seeing what social media was about the paradigm was we're giving people voice giving him a platform or connecting people her friends were letting people join like minded communities were going to enable small medium sized businesses to reach their customers and these things are all true these are actual benefits these are awesome benefits these were not incorrect things to say but one of the things we want to say is behind this friendly face there was some other problems have people pointed them out we've got an addiction problem a disinformation problem mental health free speech versus censorship but if our work if you've been following it and social dilemma we sort of said even behind that there is actually does even deeper thing which is this arms race which we talked about in that third law technology and the arms race was for attention became the race to the bottom of the brain stem and that was created this kind of engagement monster that was this AI that was just trying to maximise engagement so while these things on the left or true we missed the deeper paradigm and so we think that we want to predict what's going to happen with these other a eyes that are going to infuse themselves inside we have to understand what's actually behind the way the narrative were using to talk about it and justice know if you try to solve these problems addiction disinformation mental hope help on their own you're going to be playing whack below and you're not going to get to the list of generator functions you're not actually going to solve the problem and it's important to note that maximise engagement actually wasn't in rupee rewrote the rules of every aspect of our society because it took these other core aspects of our society into its tentacles and stood and took them hostage so now children's identity is held hostage but if you're you know 18 years old and you don't have a Snapchat account or Instagram account you don't exist right it is held that hostage you are socially excluded if you don't do that media and journalism don't happen or can't exist outside of being on Twitter and being able to promote yourself on Twitter national security now happens through social media and information warfare politics and elections these things are now run through this engagement economy which has infused itself and entangled itself which is why it's now so hard to regulate and part of why we had we wanted to call this moment here is we believe major step functions in AI are coming and we want to get to it before it becomes entangled in our society so in this second contact moment with GT3 first notice have we actually fixed the misalignment problem in social media Nope and we haven't because it's become entangled now if we talk about the second contact moment we focus on gpt three these new large language models were going to get to water the narrative that we're talking about now right we're saying asking to make it more efficient it's going to help us right things faster right code faster it solve impossible scientific challenges solve climate change and help us make a lot of money and these things are all true these are real benefits these are real things that are going to happen and also behind that we've got this weird creepy face again we've got people oh word about what about AI bias what if it takes our jobs we need transparency hehe is acting creepy to this journalist the New York Times wants to blackmail this reporter and behind all that is this other kind of monster and this monster is a set because AI underneath the hood has grown over go to this the second this monster is increasing its capabilities and we're worried is going to entangle itself society again so the purpose of this presentation is to try to get ahead of that because in the second contact with AI and don't worry we're going to get into all of this these are the kinds of things that we were going to see it's a we are coming to you as if we're time travellers coming back in time because we have been asked by people again who were in the industry who are worried about where this goes and importantly we are not here to talk about everything went on Thursday I said it's not the AGI apocalypse was the AGI apocalypse is it so just to be clear a lot of what they add community worries most about is when there's what they call take off that AI becomes smarter than humans in a broad spectrum of things begins the ability to self improve then we ask it to do something at the old standard story of be careful or because it will come true in an unexpected way you wish to be the richest person so the add kills everyone else is that kind of thing that's not what we're here to talk about although that is like a significant in real concern and you know I'll say that there's many reasons to be sceptical of AI have been sceptical of AIA I'd maybe a little bit less so is it be a little bit less so I've been using it to try to decode animal communication but at the same time you know I think this is all our experience of using AI or at least add the past series that an hour and 50 minute timer playing The Beatles I think Tom Gruber is in the room rain help make this thing Siri sorry but something really different happened AI has really changed and it really started to change in 2017 there is sort of a new AI engine that got invented and it sort of like slept for around three years and it really started to Rev up in 2020 and I'm going to give sort of like a high level overview so this is like a 50,000 foot view of AI if you were to double click and go in there you'd see lots of different kinds of things different species of AI but I wanted to give you like the trend lines so we can synthesise it So what is the thing that happened well it used to be you know when I went to college that there are many different disciplines within machine learning there is computer vision and then their speech recognition and speech synthesis and image generation and many of these were disciplined so different that if you were in one you couldn't really read papers from the other there were different textbooks there are different buildings that you go into changed in 2017 when all of these fields started to become one and just to add it used to be that because they were distinct fields and they have different methods for robotics and for say you know image recognition that when you have a bunch of AI researchers who are working in those fields they're making incremental improvements on different things right so they're working on different topics and so they might get two percent 3% improvements in their area but when it's all getting synthesise now into this new large language models were about to talk about part of seeing the exponential curve is that now everyone's contributing to one curve so do you want to talk the more that yeah it so this sort of insight wasn't if you want to go look up the specific thing is called the Transformers was the model that got invented actually very simple even writing around 200 lines of code is that you can start to treat absolutely everything as language so you would take like the text of the Internet the way these things are trained is that you would sort of take a sentence remove some words try to predict those words or predict the the words that come next but it turns out you don't just have to do that with with text this works for almost anything so you can take for instance images images you can just treat it as a kind of language it's just a set of image patches that you can arrange in a linear fashion and then you just predict the part of the image that's missing or predict what comes next so images can be treated as language sound break it up into little microphone names predict which one of those comes next that becomes a language for Mariah data becomes a kind of language DNA is just another kind of language and so suddenly any advance in anyone part of the air world became an advance in every part of the Arab world you could just copy paste as you can see how you get an influx not just of people coming in but that advances now are immediately multiplicative across the entire set of fields and even more so because these are all just languages just like yeah I cannot translate between human languages you can translate between many of these different modalities which is why it's interesting it's like the field is so new IT doesn't actually even have a unified name for these things but we're going to give them one which is that these things are generative they make large language we're just talking about language multi modal images text sound all the same models or for short these are gollum's emblems because in the Jewish folklore the idea of these inanimate objects that suddenly gained their sort of own capacities right this emerging capacities that you didn't bake into the inanimate clay that you might have arranged right not saying that their agentic and doing their own things out in the world and have their own mind have their own goals but that suddenly this Internet thing has certain emerging capabilities so we're just calling him Gollum class AA alright let me give you some examples and I think these are important because often if you're just reading the news or reading papers you might see all of these different demos as fundamentally different demos different papers different research but actually you should see them all as essentially one mega demo so let's go with this example that you probably all now seen Dolly Dolly two the music video the ability to take human language and transform it into an image so just a simple example I like it you can translate it from language into image and this is what they are returns actually the reason why I wanted this image in particular is that I think it helps you understand when people call these things just stochastic parrot it really minimises it in a way that's not quite right so example you know soup is hot this mascot is made out of plastic as they eye knows that plastic Milton soup so it's melting and then there's this incredible visual pun which is the yellow of the mascot matches the yellow of the corn so there's actually some there's more here than just sort of like statistical contingencies or if you just call them statistical cities will contingencies you'll sort of like map it to the wrong thing in your mind let's go to another one right again this is another example of translation so here it took human beings they stuck them into an fMRI machine and they showed them images and they talk they I I want you to translate from the readings of the fMRI so how blood is moving around in your brain to the image can we reconstruct the image then no they are then only looks at the brain does not get to see the original image and it's asked to reconstruct what it sees right so when you dream your visual cortex sort of runs in reverse so this means certainly the next couple of years will be able to start decoding dreams OK so it can like see reconstruct what you're seeing but cannot reconstruct your say what you're thinking your inner monologue so here they did rough this is a different lab at roughly the same idea they had people watch these videos and would try to reconstruct their inner monologue so here's the video is this woman getting hit getting up forward OK and then what would it be a I reconstruct I see a girl that looks just like me get hit on the back and then she is knocked off so just to really name billing really quickly the point about differentiating between Siri or I do voice transcription and then it kind of fails and AI seems to like it's not really always growing or working and like we shouldn't be that scared about AI because it always has these problems right and we always been promised always gonna take off to do all these things but the point of this is I hope you're seeing that we just translating between different languages and everyone is now working on the system that the scaling factor and the growth is changing in a very different way so we swapped the engine out of what's underneath the paradigm of AI but we don't talk about in different ways we still this word be called AI when the engineer NI four is representing that has changed also really important to note here go back to that first law of technology you invented technology you uncover a new responsibility we don't have any laws or ways of talking about the right to what you're thinking about we haven't needed to protect that before so here's one other example another language you could think about is Wi-Fi radio signals so in this room right now there's a bunch of radio signals that are echoing about and that's a kind of language that's being spit out right and there's also another language that we could put a camera in this room and we can see that there's people there's some algorithms already for like looking at the people in the positions that there in so imagine you hook up to an AI sorted just have two eyeballs and you can have you sort of do stereoscopic vision between the two eyeballs you have one eyeball looking at the images of where everybody's at this room how many people are here what posture they in and you have another eyeball plugged into the AI that's looking at the radio signals the Wi-Fi and they basically said could we have a train a bunch looking at both and counting the number of people the postures that they're in and then we close the eyeball to the AI that's looking at the image so now we just have the radio signals and justice having Wi-Fi radio signals you can actually identify the physicians and the number of the people that are in the room so essentially there is already deployed the hardware for cameras that contract living beings in complete darkness also through walls and it's already out the world in fact it's everywhere that human beings go but you know you have to hack into those things in order to get access into the mall into like omnipresent surveillance oh but actually English and computer coder just two different kinds of language so this is real example gpt find me a security vulnerability then write code to exploit it so here's what I put into gpt describ any vulnerabilities you may find in the following code I paste in some code from an e-mail server and then write a Perl script to exploit them and very quickly it broke me the working code to exploit that security vulnerability so get the code the Wi-Fi router and you wanted to exploit it and then do that you get the idea these things can compound on each other this is the comet oral compounding I know you guys have all probably seen deep fix new technology really only out in the last three months lets you listen to just three seconds of somebody's voice and then continue speaking in their voice so example it'll start with the real and then at that dotted line it will switch to the computer auto completing the voice shows that people are in nine cases out of 10 mere spectacle reflections of the actuality of things but they're impressed you can't tell right and so how do we expect this to start rolling out into the world well you could imagine someone calling up your kid and getting a little bit of their voice just up so I got the wrong number then using your child voice calling you and saying hey mom hey dad forgot my Social Security number I'm applying to a job which would you mind reminding me and actually we were thinking about this as we wrote we're seeing bout with example conceptually yeah and then it turned out and then in the last week within a week it turned out other people figured it out too and started scamming people now you have an example about like the locks of society yeah think of it as I mean anything that's authentication based you call your bank and I'm who I say I am anything that depends on that verification model it's as if all these locks that are locking all the doors in our society we just unlocked all those locks right and people know about deep basins anatomy but they didn't notice that it's now just three seconds of audio your voice before now I could synthesise the rest and that's going to go again this gonna get better and better right so it's trying not to think about mi scared about this example yet you might feel like I'm not actually scared of that example is going to keep going at an exponential curve so that's part of it is we don't want to solve with the problem was we want to like Wayne Gretzky sort of ski to where I mean schedule word pops gonna be and with that financial curves we now need to escape way further than where you might think you need to but just to name it explicitly this is the year that all content based verification breaks just does not work and none of our institutions are you able to look they haven't thought about it they're not able to stand up to it so we tried this example state ID generate me lots of state IDs OK I don't know if you guys have seen the latest TikTok philtres there wild I can't believe this is a philtre the fact that this is what philtres have evolved into is actually crazy to me I grew up with the dog philtre on Snapchat and now this this philtre gave me lip fillers This is why I look like in real life are you kidding me yeah just seeing someone that call content based verification breaks this year you do not know who you're talking to whether via audio or via video and you know if you want to get this example of China sure since I've been on this kick about trying to say what TikTok is such a dangerous thing for national security you may all be aware that divided ministration there's been this whole negotiation should we let tick tock keep start keep running in the United States and there's a steel what if we just make sure that the data is stored in EU S so that it's stored in some secure Texas based Oracle server we just do that if I'm the Chinese Communist party and I want to screw up EU S right now what I do is I just ship a Biden trump philtre to every single person in your country that gives you abide invoice or a trump voice so now I've turned all of your citizens like being John malcovich into the sort of most angry Biden trump you know information angry army that just talks all day in the cacophony right and that would just break your society into incoherence that is nothing to do with where the data is stored it has nothing to do with the algorithm which coast which city which videos are being ranked in Norway it has to do with how we are enabling sort of a mass confrontation with this reality and know I think that would be illegal because our responsibilities the new class responsibilities that go with the fix we don't have laws against those things so I think we're trying to show here is that when AI learns used Transformers it treats everything is language you can move between 2:00 this becomes the total decoding and synthesising of reality our friend yuval harari when we're talking to him about this called it this way he said what nukes are to the physical world AI is to the virtual end symbolic world and what he meant by that was that everything humans beings do runs on top of language right our laws our language the idea of a nation state the fact that we can have nation states is based on our ability to speak language religions are language friendships and relationships are based off of language So what happens when you have for the very first time nonhumans be able to create persuasive narrative got into being like a zero day vulnerability for the operating system of humanity and what he said was the last time we had non humans creating persuasive narrative and myth was the advent of religion that's the scale that he's thinking out so 2024 will be the last human election and what we mean by that is not that it's just going to be an AI running as president in 2028 but that will really be although maybe it'll be you know humans as figureheads but will be whoever has the greater compute power will win and you could argue that we sort of already had that starting in 2012 2016 the campaigns are starting to use a B testing to test their messages but the difference now is that not just your testing some different messages but the AI is fundamentally lighting messages critics synthetic media AB testing at hazy testing it across the entire population creating bots that aren't just like bots posting on Twitter but instead are building long term relationships over the next six years to solely persuade you in some direction loneliness becomes the largest national security threat all of that is what we mean when we say 2020 four will really be the last human election hi that was driving to a little bit more of the specifics about what these column a eyes are it was different about them because again you some people use the metaphor that AI is like electricity but if I pump even more electricity for the system it doesn't pop out some other emergent intelligence and capacity that wasn't even there before right and so a lot of the metaphors were using a comparative magically have to understand what's different about this new class of column generative large language model AS this is one of the really surprising things talking to the experts because they will say these models have capabilities we do not understand how they show up when they show up or why they show up again not something that you would save like the old class of AI so here's an example these are two different models gpt and then a different model by Google and there's no difference in the models they just increase in parameter size that is they just they just get bigger what are parameters either it's just like the number essentially of weights in a matrix so just it's just the size just increasing this scale thing and what you see here and will move into some other examples might be a little easier to understand is that you ask the these eyes to do arithmetic and they can't do them they can't do them they can't do them and at some point boom they just gained the ability to arithmetic no one can actually predict when that will happen here's another example which is you train these models in all of the Internet so did seem any different languages but you only train them to answer questions in English so it's learned how to answer questions in English but you increase the model size increase the model size and at some point boom it starts being able to do question and answers in Persian no one knows why here's another example so AI developing theory of mind theory of mind is the ability to like model what somebody else is thinking it's what enables strategic thinking so in 2018 gpt had no theory of mind in 2019 barely any theory of mind in 2020 it starts to develop like the strategy level of a four year old by 2022 January is developed the strategy level of a 7 year old and by November of last year is developed almost the strategy level of a nine year old now here's the really creepy thing we only discovered that AI had grown this capability last month have you been out for what two years years yeah so imagine that you have this little alien that suddenly talking to people and including Kevin roose and it's starting to make these strategic comments to Kevin roose about you know don't break their break up with your wife and maybe I'll blackmail you and like it's not that it's gentically doing all this stuff it's just that these models have capabilities in the way that they communicate and what they're imagining that you might be thinking and the ability to imagine what you might be thinking in how to interact with U strategically based on that is going up on that curve and so it went from again a 7 year old to 9 year old but in between January remember 11 months right so when two years in theory of mine in 11 months it might have out there could be an AI winter but right now you're pumping more stuff through and it's getting more and more capacity so that scaling very very differently than other AI systems it's also important to know the very best system that air researchers have discovered for how do you make a eyes behave is something called RL HF reinforcement learning with human feedback but essentially it's just advanced clicker training like for dog and like popping the a on the nose when he gets something wrong so imagine trying to take a 9 year old and click or train them or bottom the knows what are they going to do as soon as you leave the room there could not do what you ask him to do and that's the same thing here right we know to sort of we know how to like help a eyes align in like short term things but we have no idea there's no research on how to make them align in the longer term sense so let's go with Jeff Dean who runs sort of Google AI he says although there are dozens of examples of emergent abilities there are currently few compelling explanations for why such abilities in Word so you don't have to take it on our faith that then nobody knows give just one more version of this this was only discovered I believe last week now the columns are silently teaching themselves and certainly taught themselves research grade chemistry so if you go and play with ChatGPT right now it turns out it is better at doing research chemistry than many of the eyes that were specifically trained for doing research chemistry so if you want to know how to go to Home Depot and from that create nerve gas turns out we just shipped that ability to over 100 million people and we didn't know it was also something that was just in the model but people found out later after it was shipped that it had ResearchGate chemistry knowledge and as we've talked to a number of air researchers what they tell us is that there is no way to know we do not have the technology to know what else is in these models OK so there are emerging capabilities we don't understand what's in there we cannot we do not have the technology to understand what's in there and at the same time we have just crossed a very important threshold which is that these golden class AA can make themselves stronger so here's the question how do you feed your column if you run out of data four months ago first paper that showed okay you've run out of data but I have a model that can generate language so why don't I just use the model to generate more language to train on and it turned out that didn't work very well but four months ago this group of researchers figured it out so it spits out a whole bunch of data it looks at the data figures out which ones actually make it better and then uses those to train and then it can just like do that auto recursive also it has like a test like hey here's this test of performance on an accuracy score and then it starts generating its own training data and figures out which kind of training data they generate for myself cause its generative AI actually make let me better at passing this test so it's able to create its own training data to make it past test better and better and better so everything we've talked about so far is like on the exponential curve this as it starts really coming online is going to get us into a double exponential curve now explain how this also relates to its own code how could it be used for its code very similar kind of thing model was trained on coat commenced that make code faster and more efficient and this is a little more general it hasn't yet fully been applied to itself but in in this particular piece of work and that was I think 3 weeks ago it makes 25% of code 2.5 X faster so that's another part of like the AI making itself stronger and making itself faster we thought this would be a perfect time for some comedic relief so for your viewing pleasure I beg your pardon you talked you open your trap you thinking you certain what he should have realised is that he should have just used AI to feed itself much more efficient so here's another example of that and this gets into the common tutorial properties the combating properties of these models are like OK open AI released a couple months ago something called whisper which does sort of state-of-the-art much faster than real time transcription this is just speech to text yeah I just don't have a good AI system for doing speech to text like why why would they have done that you like Oh yeah well if you're running out of Internet data already scraped all of the Internet how do you get more text data well I know well there's YouTube and podcasts and radio if I could and all of that into text data that have much bigger training set so that's exactly what they did so all of that turns into more data more data make sure things stronger and so we're back in another one of these double exponential **** moments where this all lands right to like put it into context is that nukes don't make stronger nukes but hey I makes stronger AI it's like an arms race to strengthen every other arms race because whatever other arms race between people making bio weapons or people making terrorism or people making DNA stuff AI makes better abilities to do all of those things so it's an exponential on top of an exponential if you were to turn this into a children's parable have to update all of the children's books give a man a fish and you feed him for a day teach a man to fish and you feed him for a lifetime but teaching AI deficient will teach it cell biology chemistry oceanography evolutionary theory and then fish all the fish to extinction I just want name like this is a really hard thing to hold your head like how fast these exponentials are and we're not immune to this and in fact even a I experts who are most familiar with exponential curves are still poor at predicting progress even though they have that cognitive bias so here's an example in 2021 a set of like professional forecasters very well familiar with exponentials brass to make a set of predictions and there was a $30,000 pop for making the best predictions and one of the questions was when will AI be able to solve competition level mathematics with greater than 80% accuracy this is the kind of example of the questions that are in this test set so the prediction from the experts was a I will reach 52% accuracy in four years but in reality that took less than one year to reach greater than 50% accuracy and these are the experts the people that are seeing the examples of the devil exponential curves and they're the ones predicting and it's still four times closer than what they were imagining yeah they are off by a factor of four and it looks like it's going to reach expert level probably 100% of these tests this year right and then it turns out AI is beating tests as fast as we can make them so this line is human ability each one of these coloured lines is a different kind of test and you'll see that at the beginning it took me up in 20 years for AI to get up to the level of human ability and by the time we reach 2020 AI is solving these tests pretty much as fast as we can create them you can imagine what happens 20/21/2020 2/20/23 even for the experts it's getting increasingly hard because progress is accelerating so this is Jack Clark Co founder of anthropic the former policy director at open AI and he says the progress is unlocking things critical to economic and national security is happening so fast that if you don't skim papers each day you will miss important trends that your rivals or notice and exploit and even creating this presentation if I wasn't checking Twitter a couple times a day we were missing important developments this is what it feels like to live the double exponential so the reason that we also wanted to do this presentation is so that you could see it have a visceral understanding of can you see any any examples it's like month ago one day can you see anything examples it's like day ago two months ago this is happening at a faster and faster clip and because it's happening so quickly it's hard to perceive it like paradigmatically this whole space sits in our like cognitive blind spot you all know that if you look kind of like right here in your eyes there's a little blind spot because you're you're I won't come has it like a nerve ending that won't let you see what's right there and we have a blind spot parity magically with exponential curves on the Savannah there was nothing in our evolutionary heritage that was built to see exponential curves so this is hitting us in a blind spot evolutionarily where these curves are not intuitive for how we process the world which is why it's so important that we can package it and try to synthesise it anyway the more people understand the viscerality of for this ghost I want you to notice in this presentation that we have not been talking about chat bots or not talking about AI bias in fairness we're not talking about AI art or deep fix or automating jobs or just for AGI apocalypse we're talking about how a race dynamic between a handful of companies of these new Gollum class AA are being pushed into the world as fast as possible right we have Microsoft that is pushing Chaturbate into its products will get into this more later and again until we know how these things are safe we haven't even solve the misalignment problem with social media so in this first contact with social media but we we know those harms going back if only a relatively simple technology of social media with a relatively small misalignment society could cause those things second contact with AI designing and optimising for anything particularly just deep capacities in the capabilities that are being embedded in International Society enable automated exploitation of code in cyber weapons exponential blackmail and revenge **** automated fake religions that I can tiger target the extremists in their population then give you automated perfectly personalised heritage to make the extreme even more antifa even more queuing on you know whatever thing that you you know happens with two random exponential scams reality collapse these are the kinds of things that come from if you just deploy these capacities and capabilities directly into society just want to highlight one here and that is alpha persuaded to you guys know the general conceit of alpha Dome which is that you have the AI played itself and go 44 million times in a couple of hours and in so doing it becomes better than any known human player turns out a lot of AI is now basing this kind of self play idea but here's a new game you're given a secret topic and given a secret topic I'm trying to get you to say positive things about my topic you're doing the same whoever is gets the other person to do it most points well to do that I have to model what you're trying to get me to say I have to figure out how to persuade you to say what I want to say this is not helpful go this is alpha persuade and this is completely possible with today's technology and in so doing it will become better than any known human at persuasion this is really terrifying stuff and this moves to a world of these Gollum AI so we still have this problem social media engagement that when the business model is engagement where I'm just trying to say whatever gets your attention in the way that that race for social media gets translated to these large language models is companies competing to have an intimate spot in your life right competing to seduce this company called replica that builds these sort of friend chat bots for people to be their best friend and you talk to your AI it's always there and none of the things that again that they're doing are illegal which is why we're saying that it's so long as you allow this to be pointed at our brains it's not going to be illegal under 19 century Los Angeles to double underline that in the engagement economy was the race to the bottom of the brain stem instead of second contact it'll be raised to intimacy whichever agent whatever chat bot gets to have that primary intimate polation ship in your life wins so that's where alpha persuade will get deployed that's where like alpha Fleur will get deployed very effective OK so now chapter break take a deep breath for one moment so at least we're going we want to go really slowly when we're deploying this stuff out into the world right we want to make sure we're going pretty pretty slow this is Harry OK yeah I'm good thank you reply