Augmented Intelligence

[Music] it's it's always a an honor and a privilege to stand on this famous stage and bring hopefully some interesting and debate worthy science for all of us this evening so quantum black we are machine learning and artificial artificial intelligence outfit we try to bring together data science software engineering technology design and even anthropology to solve complex problems using artificial intelligence and we do that in areas like pharmaceutical around drug discovery and development and commercial in advanced industries like aerospace and so on in financial services we started life in Formula one and we still try to bring that edge of performance to the work that we do and through all of that we're big fans of augmented intelligence we like to talk about augmented intelligence rather than artificial intelligence because we're big believers in bringing the human and the Machine perspective together and we love this metaphor of the Ironman suit we like to try and build technology which gives humans superpowers rather than replaces them and there are different views on this as this field evolves very rapidly and just before introducing our speakers and I want to just show two short videos which will hopefully work if the if the gods of technology are with us one showing the something positive and exciting hopefully and one showing something potentially scary so if we can maybe try and show the first video of the conversation here we go this is just go back to this example let's say you want to ask Google to make you a haircut appointment on Tuesday between 10:00 and noon what happens is the Google assistant makes the call seamlessly in the background for you so what you're going to hear is the Google assistant actually calling a real salon to schedule the appointment for you let's listen [Music] I'm happier hi I'm Kyle women care capture our client I'm looking for something on May 3rd so I give me one second mm-hmm or what time are you looking for well at 12 p.m. we do not have a quality available but closest we have to that is a 115 do you have anything between 10:00 a.m. and 12:00 p.m. depending on what service she would like what service is she looking for just a woman's haircut for now okay we have a 10:00 o'clock 10:00 a.m. is fine okay what's her first name the first name is Lisa okay perfect so I will see we thought it's 10 o'clock on May 3rd okay please thanks Lee how do they pay back very exciting deployment of these kinds of technology this was just about a week ago I think that the Google announced an and on the flip side there are also areas where these same kinds of technologies around machine vision natural language processing natural language generation can seem intimidating so I was recently in China and someone said to me well you realize that everywhere you go you will be recognized by the CCTV cameras found this was interesting and it turned out that actually a BBC journalist has has gone and investigated this recently and this is a second video which is made up of two clips so bear with us while we switched between the two clips well let's see how long it takes you to find me thank you very much let's go [Music] they heads out into the streets and then we will pick him up as he goes into a railway station yeah it sounds good it's full of Japan I told to tell you this up you pay how about what cheers whoo biceps you pay I also the t-90s who's everyone same ship it just starts unzip pull Dajjal hello Nancy but I did middle so hot you sir on the hoodie hunger to learn you can see well here we are then I've just got out of the car close to the city center and for the purposes of this exercise the plan is for me start walking in the direction of the bus station my image has already been flagged to the authorities as a suspect and in theory it should only be a matter of time [Music] so already on this bridge I can see one two three CCTV cameras of course there's no point hiding from them just keep on walking right behind me you can see just over over my left shoulder there hello guys I've been expecting you maybe these guys aren't in on the joke two interesting perspectives they raised a whole number of questions around ethics around deployment around the way that these kinds of technologies interact with us in our lives and so on and so this evening we have three different perspectives on this the first is from Martha who's the data scientist and and practitioner in this area who's going to talk about approaches to bringing humans and technology together to solve problems and how we can think about that and their questions it raises will then have Corinna who is a philosopher and we'll think about how augmented intelligence can be reflected in how our mind our consciousness is also very much present in the tools and sort of external sort of things that we use every day and finally James who will think about what does augment intelligence mean for us as humans and how can we to some extent augment the way that we interact with technology as society in our behaviors evolve over time so with that please welcome our first speaker who's Martha from Connor black [Applause] good evening how do you feel about algorithmic making important decisions about about your life what if for example you sat at the back of a self-driving car how would you feel and would you be comfortable if in the future instead of consulting your GP you had to consult an algorithm what if this algorithm suggested that you should have surgery or that you should change your medication would you trust it artificial intelligence is shaping our lives in ways that we have only now started grasping and so I'm going to share a perspective about what our symbiosis with intelligent systems may look like in the future and also discuss how we are likely how can we make that symbiosis safe and how can we make it for the benefit of humanity so I'm a little scientist at quantum black as Chris said we work with clients using data analytics machine learning artificial intelligence techniques to solve complex business problems and it is a multidisciplinary effort combining skills as data science software engineering or design so first of all let's see what we mean by artificial intelligence artificial intelligence exists as a field of computer science for a long time arguably since the days of Alan Turing and formally as a concept since the 50s however the very big technological breakthroughs that we have observed in the past few years voice recognition systems or image recognition systems or machines that can beat human champions ingo or chess all these come from a particular branch of artificial intelligence that is called machine learning in machine learning the differentiating factor is that systems can learn from experience essentially from data without being explicitly programmed in other areas of artificial intelligence a human still has to program and to essentially instruct the Machine about how to solve the problem and then maybe the algorithm mimics intelligent behavior but machine learning is slightly different in that respect because there what we only give to the algorithm is a learning strategy and then the algorithm is using that to figure out complex relationships in the data and solve the problem and so potentially it can solve problems that that we couldn't solve so when we hear about artificial intelligence we hear things in the news such as in the next 10 to 15 years artificial intelligence is going to replace almost 1 million jobs in the UK public sector alone however I want to challenge a little bit thinking of artificial intelligence as a form of substitute as a substitute to human intelligence because essentially the two are quite different qualitatively humans we have common sense we have imagination we have intuition we are able to think rationally but we are not always rational we're also emotional artificial intelligence systems on the other hand have none of these properties however they are much faster at making computations compared to humans they are also much more efficient at pattern recognition but at the same time they only limit themselves to learning from data so and so sometimes we think of artificial intelligence systems as autonomous agents or as a concept of singularity of of an agent that can transfer their knowledge on many different respects and so complex problems but in reality they don't have the same adaptability as humans do they still rely on humans to be able to to sort to be able to be given an objective and solve a problem and so an alternative conceptualization of artificial intelligence is what we call augmented intelligence which is what we're going to talk about today essentially as a space where human intuition coexist with the efficiency that artificial intelligence techniques can provide achieving performance beyond what human intelligence or all machine intelligence can achieve on their own and to achieve augmented intelligence it would be good to also have a perspective on what tasks that right now humans do are better to be left to humans and what other tasks are potentially better to be taken on by machines for example humans are much better at writing a short story telling a joke or designing a new product on the other hand we have seen machine learning systems being a lot better at for example playing go playing a certain other games or even at driving a car or flying a drone so let's see how human intelligence compares with artificial intelligence looking at an example at prediction prediction is the most common task that machine learning system typically solves humans however we are also good at prediction arguably a lot of the technology that we have built since prehistoric times come from the fact that we are able to predict the future and anticipating and create the artifacts it will protect us from it so who is better at prediction humans or machines let's see an example so this is a case study from some work that we've done at quantum black with from a suitable client and the task is very simple sales forecasting what the client wanted they they used human analysts it took a very long time to produce yearly forecasts and essentially they wanted to make the process more efficient but also they wanted to increase accuracy so when it comes to efficiency machine learning system is perform that a lot better than humans so for that one we used space state-space models as well as Bayesian models and then we optimize over thousands of parameterizations and all this can be achieved by machine learning system in a few hours if you run it on your laptop a machine a human in comparison couldn't perform all these computations but still to produce a final forecast it would take several weeks if not months however when it comes to accuracy we see something different we see that actually there are certain types of markets and certain types of products where humans are better than machines and others world machines are better than humans so essentially we see that they do perform across the board across all forecasts about the same and why is that this is because if you have if you imagine a time series in a financial market or any or any time series that is that follows a very clear pattern that essentially has very strong seasonality or that you have a lot of events that recur many many times then they machine in those cases is able to learn and it is actually able to be more precise in the forecast it will make compared to human however there are always these Black Swan events there are for example often changes in the economy or unusual competitor activity essentially there are events that render all the data that you have useless and because the future looks like see nothing like the past and it is in those cases with human intuition they innate ability that humans have to hear something and use that knowledge to transfer it to complete a different domain that enables us to be able to deal with these events better than machines can do and essentially in that particular case the solution to improving overall forecast accuracy came from being able to identify which were the forecasts that the machine was most suitable for and which were the forecasts were the humans were most suitable for and essentially how that works in the practice is that there is an automated forecasting process it returns it returns its as its processed along with the uncertainty and where we see a lot of uncertainty where we see that the machine really is struggling there we can have a human reviewing it or even modifying the output of the forecast so these intelligent systems are going to be more and more present in our lives as augmented intelligence or even as machines solving individual problems that we are used to have them being solved by by humans and some of machine learnings greatest achievements involve Dusk's where they really they really demonstrate better performance than humans for example being better than humans than the human champion are playing go we also have right now systems that are better at diagnosing certain types of tumors compared to human doctors we have also started seeing some evidence that artificial intelligent systems might be better at flying drones compared to human pilots so we've seen all these but at the same time are we giving away too much of our decision-making power to machines as we see that because at the same time there are also negative examples of using artificial intelligence none of these were intentional but they have happened anyway so for example recently uber suspended their self-driving car program after a car killed a pedestrian while they were testing it this was a disaster and of course we don't want such events to happen or for example Google their sentiment analysis system assigns a negative or useless sign and negative sentiment to terms such as homosexual or Jew similarly Amazon when they were trying to extend their free say prime same-day delivery membership they used the machine learning system to identify customers to whom that should be extended to and they ended up select they ended up not selecting people that were in minority neighborhoods and of course none of these was intentional I mean Amazon for example weren't training their systems to exclude pay to exclude people on the basis of their ethnicity however the system itself ended up having that output that was an intent they intended but but still potentially harmful to people so and this is another key difference that humans have from machines essentially artificial intelligence systems they optimized for a single thing they optimized for performance they optimized for accuracy however humans to live in society we need we need other concepts we need for example resilience we need to know that the tools that we use are safe for us and for others we need transparency we need to be able to justify to other people why we are making the decisions we make we also need fairness and this is a big concern because the world in the past hasn't been perfect but as these systems are trained from data that come from the past how do we make sure that they don't propagate the same bias isn't the same unfairness and has existed in the past it is possible to create systems that actually mitigate some of those risks and this is active work that we are doing with quantum black and we have seen an increased interest across several different industries for example banking or pharmaceutical companies and it's a good thing that people are getting concerned with these ideas of course so the first design principle is resilience we need to try to align algorithms to our objectives and we need to know that when we test them in the wild in real-life scenarios they're going to behave according to expectation and this means developing the monitoring mechanism developing detection mechanisms and also developing corrective mechanisms for when things go wrong and this is a joint responsibility it involves designing systems that are robust and resilient and even writing algorithms a test that can test themselves but at the same time it involves defining within businesses within companies and within society in general who are the accountable parties who is going to respond to be responsible about systems and its failures the second concept is explained ability or interpretability which is an active area right now in machine learning a common criticism to machine learning systems is that they are black boxes in reality they are not black boxes exactly because we have access to what the output will be to every input but still the models that they compute they tend to be complex hardened to hard to explain in human language and so there is an area right now in in artificial intelligence that tries to explain the outputs of of black box models and what you see there is an example of certain of such a method with this one is called lime there are many others these methods essentially exploit locality they exploit the fact that even if a model is a black box and even if it's nonlinear the area around a particular individual sample is going to be approximately linear and so using that property you're able to generate individualized explanations so for example and this is the example that you see on there on the page if for example we are writing system for diagnosis to diagnose a patient the system can generate explanations as to why we got this diagnosed so for example if the diagnosis is you have the flu it can generate okay what were the symptoms that the system identified to that particular patient and this can help make it more transparent especially if we if you have customer facing applications the third concept is fairness how do we design systems that are fair there are two key concepts around fairness the first one is protecting attributes essentially making sure that there is no sensitive information that could potentially be the basis of discrimination somehow leaked into the training data and the second one is making sure that the output of the algorithm are fair for the different groups that might be involved for example if we have a voice recognition system we want it to be equally good at recognizing the male or the female voice or if we have a face recognition system we want it to be able to recognize faces with the same mount of accuracy regardless of ethnicity of the person and so to conclude let's see what we what we saw today we saw that artificial intelligence is not a substitute of human intelligence and that the two can work together to create augmented intelligence to improve performance we also saw that the key to augmented intelligence is to be able to see which tasks are best left to human and which others are best to be taken on by machines the third thing we saw is that it is a very exciting future what artificial intelligence is promising us but at the same time we need to make sure that the systems that we design are aligned to our goals and values and principles as humans and this is of course the responsibility of all of us the people who are designing those systems but there is an overall responsibility for everybody to be informed about what this change means and to also be actively involved in the debate to guide how we should shape our symbiosis with intelligent systems I asked at the beginning how do you feel about this change that is happening but I would like to invite you to think what will your role be in this change thank you very much [Applause] thanks Martha so there we have the the practitioners view from the from the frontline and hold that one in your mind as we shift gears to the the philosophers view Karina so thanks it's really really great to be here and so as a philosopher I spend a lot of time thinking about human cognition and what I want to do today is to present to you a sort of metaphysical view about how the technology that we're using and is becoming integrated into our cognitive capacities so really augmenting our intelligence and what that means for data for our privacy and for things like thought manipulation okay so let's begin by considering a case that made a lot of headlines so in 2016 there was this encryption dispute between Apple and the FBI and in this case the US Central District Court of California ordered Apple to enable the FBI to unlock the iPhone belonging to one of the perpetrators of a mass shooting in San Bernardino County and as I'm sure many of you know okay this like I said it made a lot of headlines and in fact Apple refused to do this and I think in doing so it gained a lot of support from the public so it was a good a good move I think by the CEO in part because they they realized that it was going to threaten data security of other users right so they wanted to stop the right kind of precedence and now the FBI actually found their way into the phone anyways and but I think the case sort of violated a lot of our serious concerns about how our very personal data is going to be protected or not protected by the law and actually since then the FBI has used the dead fingers of corpses of mass shooters to try to unlock iPhones using the fingerprint recognition technology they were unsuccessful so you can but still it I think it's so I think there's something kind of disturbing about that and maybe not everyone agrees and but I want to kind of articulate what I think is is sort of wrong there and so I think part of this has to do with the fact that our democratic constitutions at least most of them explicitly protect our right to mental privacy and they explicitly protect our right to freedom of thought right and this has typically amounted to things like brain protections so we are protected from being mandated to use neuroimaging lie detector tests and we're protected from having to take like chemical intrusions that would meddle with our mental functions and and part of the reasons for this are these explicit protections of free free thinking so you can't be forced to take drugs unless there's a real medical justification for that and but currently okay the law does not extend these types of protections to our devices so in fact I want to suggest that the law is sort of based on this outdated cartesian conception of where where our mind is so this cartesian legacy is that our mind is just nothing more than our brains so in a sense it's your mind is locked in your skull right so you need nothing more than the brain to explain everything that's going on in the mind and a result of that commitment is that the mind is sort of impenetrable by its very nature and I think actually in coggan science a lot of people are pushing back against this picture so the idea now is that our mind is much more fragile than that and that it's easily manipulated and nudged and influenced by external factors so this has a sort of broader picture of what the mind is and so this is what I kind of want to convince you of today that your mind is not just your brain that your mind is more than your brain and in fact I want to convince you of something a bit like this and this is the picture I want to paint and of course it's a bit dramatic right but it's not a metaphor so I really want to convince you that your mind is merging and with the technology you're using so in other words the technology that you're using is becoming so functionally integrated into your cognitive capacities that it's on a par with your brain on a par metaphysically speaking but also I think ethically and maybe even sort of legally we should start thinking about these things as similar ok ok so I could think a good sort of safe place to start here is to ask the question what a fact is technology having on our cognitive capacities and so as a philosopher this is a question that we've been dealing with for thousands of years not that we have any solid answers ok but I'm going to I'm going to argue for one tonight so the first thing to say is that Socrates actually had a position on this so Socrates thought that technology would have a diminishing effect on our memory capacities and on our social skills so he actually argued against the major technological shift that was happening at his time which was a shift from the oral tradition so from orosi to literacy and Socrates actually warned against the dangerous consequences of writing words down okay now I don't know about most of you but I actually think literacies a good thing and I think overall it's had a net positive effect so um I don't think it's a crazy position though I think I should highlight that so I think actually we still hear kind of similar line today right so you you often hear people say that our use of technology and our reliance on things like social media for example is making us forgetful and sort of turning us into a social morons okay so I think this picture is wrong so I'm gonna argue that actually technology is having an augmenting effect on our cognitive capacities that's not diminishing it's augmenting us it's making us smarter it's enhancing our intelligence and I'm even gonna argue that this is a part of who we are as humans to use technology so we ought not be afraid of it okay so let's consider a study here what effects is Google having on our memory so this study showed that our reliance on Google is not dumbing us down as much as it is reshaping our capacities so reshaping the kind of things we remember so in this study and Betsy Sparrow and her colleagues and found that when participants knew that certain information sort of relevant information would be accessible to them in the environment they were less likely to remember that informations about particular information and they were more likely to remember how and where to access it okay so it's it's changing the kind of things we remember I think this should ring very true to us right so how many people in this room still remember their best friend's phone number I don't even remember my phone number half the time I move around a lot and so that's in part because we know how and where to access that information right it's in our phones and we trust it we feel like it's pretty safely secured there and I think that's true of a lot of the information that we used to have to remember so you could say that your your smartphone and your various devices are taking over a lot of the memory functions that that your brain wants sort of had to do and I think this is a good thing because it's freeing up our internal resources to do the kind of things that humans are really good at so things like more moral reasoning things like empathy understanding creativity even sort of context specific relevance determination which computers are very bad at so lots of things that we're still good at the computers aren't ok ok so our best science tells us that the mind is just an information processing system and cognitive science and computational neuroscience are committed to what's called the computational theory of mind the computational theory of mind says that thinking is just a process of symbol manipulation which is typically carried out by the brain in the form of neural computation ok but a lot of cognitive scientists are starting to suggest that in fact it might not just be the brain that's doing the relevant information processing for us so it could just it could be now a conjunction of both artificial and biological resources working together that explains a lot of our intelligent capacities so what I have in mind here is not just things like smartphones and really good devices but also things like simple technology so the example of writing so a pen and paper so the Nobel prize-winning physicist richard fineman described his thinking is happening with his pen in his notebook right that's how he described it from his perspective and I think you'll hear a lot of expert mathematicians say something similar so they'll expert mathematicians need to have a pen and paper in order to perform their complex calculations so they create external symbols and then they push them around on the page in order to come up with different serums and actually there's kind of an old joke that all the mathematician needs to do her work is a pen and paper and a wastebasket and all a philosopher needs as a pen and paper philosophy jokes typically don't get that kind of laughter thank you that's charitable thank you very much okay so um Herbert Simon and the fathers of computational of computer science and the computer and defended this principle of near decomposability which has now been sort of accepted as a tenet of the cognitive and brain sciences and according to this principle it doesn't matter where some part is located and it doesn't matter what that part is made of okay what matters is the rapid and intensity of the interactions between a particular part and the overall cognitive system so we can sort of measure you know what sort of dimensions matter here so the typical candidates are things like bandwidth connections effective information flow and by directionality of connections so having a two-way connection so music and sorry um so what matters is the intensity of the connection so there's a bi directionality between them and that they're a stable connection and so what I what I would suggest here is that for many of us as we know we don't use all the parts of our brain right there's often parts of your brain that get unused at least not very regularly - but a lot of us use our phones all the time right this is a constant connection the information the phone is highly accessible to us and that we rely on that information to effectively guide our behaviors and we endorse the information is true so we don't question its porosity and all this would be sort of similar to what's going on in other parts of your brain and it in fact there's even studies that show that so if most of us look at our phones within the first 15 minutes of waking up all right and that we some of us even experience sort of phantom ringing and phantom buzzing and so it's all this is sort of suggests a sort of very tight connection between you and your phones so the suggestion I'm making is that for our science of the mind and so far as the science of the mind is concerned the skin in the skull are not theoretically significant boundaries okay and this is a claim that was defended by philosophers Andy Clark and David Chalmers and Clark in particular argued that humans are the offloading ape so we're the ape that likes to offload our dirty cognitive labor into the environment so we have machines and tools that do are the things we don't want to bother doing and we're very good at that so this is sort of like one of the core ideas of his philosophy and there's this nice New Yorker piece actually just came out a few weeks ago and there was all about Clark's Clark's ideas so I highly recommend it another one of his sort of core ideas is that language plays this key role in into our development or cognitive development so language is an external symbolic system which we created and which allows us to have higher-order thoughts so thoughts about thoughts beliefs about desires second order third order so on ok so what does this mean okay so what does this tell us about you know privacy and data so on what can we learn from these ideas crazy philosophical ideas hmm ok so first of all I want to consider some applications in healthcare and so the case of sort of a population of Alzheimer's sufferers in inner-city st. Louis so they scored dismal e on some cognitive memory test scored really poorly on them so poorly that doctors thought based on their scores alone they ought to be relocated into full-time care facilities um but doctors were surprised because they actually seem to be functioning very well in regular life even an inner city life which can be very complex places extra demands urban life places extra demands on the brain and so what they did was they visited their homes and what they saw was that these patients had restructured their home environments so they had labels all over their walls about what to do you when to do it if family photos with people's names written on them ok names of friends and family so that they didn't have to rely on their internal memory right so that information was in the environment when they needed it some of them had even taken doors off of their cabinets so they created open storage cabinets that they didn't have to remember where their checkbook was or were the plates and and cops were so for sort of creative strategies too so if you think well if we had removed these people from their homes surely they're calling it a functioning would have dropped off right so what that suggests is that there might be ways in which we can use especially with technology or devices to sort of fill in the cognitive gap so people who have neurodegenerative conditions like in this case might be able to rely on things like personal devices so the devices that can then when to take pills or order groceries for that or you know even use a visual recognition technology to remind them who's who in their life I think there's a nice sort of lesson we can take from this case so you might think well nothing I've said is really all that new right so I mean advertisers have known for a very long time that they can hijack our attention and influence our decision-making and that's something they're very good at and in a sense you're right so this is exactly the type of lesson that has influenced the philosophical view that I'm describing and but I think what's new here is that with some of the technologies that we have now so that the things we carry around all the time and that are online our thoughts can be intruded upon even before they're fully developed so even before you sort of come to a conclusion on a particular topic you might be getting influenced by you know advertising in your social media stream for example so there's a sort of double-edged sword here so there's a great power in being able to use these technologies but I think there's also a big concern that they might be used in the wrong way and especially I think with big data so big data and now can nudge people in a very cost-efficient way so they can sort of customize based on your data a nudge that's going to be good for you I know just it's gonna be effective for you and they can track how effective that is with a relevant algorithm that learns based on the user's behavior and then it can adjust according to what kind of results it's getting this is a whole other level of just the kind of nudging we've traditionally seen by government by business and it raises this sort of whole branch of thorny ethical questions I sometimes sort of joke that there's a new branch of applied ethics that I call Facebook at it it opens up all these new questions about when does a nod you become manipulation right and can we really you know what is the ethics of trading your own personal data and how should the law protect your data so these are some of the big questions that I think philosophers and industry have to grapple with now and I think also actually the public so we need to think about how we feel about these these new technologies and and part of that is of course understanding how they work exactly okay so I actually wrote this article earlier this year where I kind of confessed that my phone would tell a more intimate story about me than my best friend so I just want to highlight just how much data is on these devices right so your phone has higher quality and quantity of information about you than any piece of hardware in the history of mankind including your brain okay my phone knows who I talk to when I talk to them what I say what they say back how long it took them to reply right where I was when I said those things it has all my photos all my purchase history all my biometric data it has my notes to myself and it has all of this dating back years that's and it's with high accuracy so our brains are susceptible to false memories as I'm sure many of you know and so the phone records all about it with a very high degree right it's not changing or altering information so I mean so again this this suggests that there's something very intimate and different going on with these technologies and actually on the US Supreme Court Chief Justice John Roberts actually cited this observation okay in 2014 the observation that your cell phone holds a higher quality and quantity of information than any piece of hardware before in history for the courts in a court's decision so in his written opinion of a court's decision justifying why the Supreme Court now demands that police get warrants before they search your smartphones in searches incident to arrest which our particular case in the u.s. so there's a new precedent now that other objects can be searched but when it comes to smartphones and devices in the u.s. and police have to get a warrant even if it's when you've been arrested and that's the reason he's given for that and in that justification he also notes that these phones and these devices have become such a sort of persistent part of daily life that he says the proverbial visitor from Mars would likely conclude that they're a part of our human anatomy so I think many of you are probably clutching your phone right now as I say that right and it's not it's not a crazy thought so it's not only a kind of extended mind it's a kind of extended body as well and so what I want to suggest and what some philosophers are beginning to suggest is that as a result of this interconnectedness the phone and any sort of device that contains this kind of high amount of personal data ought to have the same kind of protections that were traditionally given to your brain and ethical considerations okay so I'll just end quickly so of course we can't go back to being a literate right so even if we wanted to but I think we can still take some nice wisdom from Socrates so Socrates says to know thyself okay and I think it's important to remember that we are the offloading aids so this is something that we do very well and what we want is to create a world in which everyone feels comfortable and safe actually offloading this information in the ways that we will do and will continue to do and what I would commit is that in 2018 to know thyself is to know thyself own thank you [Applause] thanks comunist some some mind-expanding concepts they're quite quite literally and I'd now like to introduce our third speaker James Hewitt who will help us think through how we can augment ourselves as we go through all of these changes thank you Chris what a fantastic series of presentations I was standing there in the background while I was trying to fit this headset listening in intently give me a lot of ideas to think about my name is James Hewitt and I'm a performance scientist I research human performance in both sports and in business I'm particularly interested in sustainable high performance and how lifestyle and work interacts to maybe enhance or inhibit that performance and this evening I'm going to explore some of the same themes that have been talked about already but perhaps through a different lens now we've worked for quite a long time in high-performance environments and I work with a company called hints at performance we have worked in Formula One for over 20 years we often talk about Formula One as our laboratory we've been quite successful in that context actually twelve of the drivers that we've worked with the one Formula One World Championships we've got a great success rate there and often I talk about Formula One as an example but how many people here have driven a Formula One car not many you know most of us we don't have the luxury of Formula one focus you know I think the many of us here we'd probably find that the work we do likely many of us and knowledge workers use our brains to generate our value we find that that knowledge work is becoming increasingly complex there's a lot to pay attention to a Kareena alluded to that in terms of how we use and engage with our smartphones you know 79 percent of people according to some research look at their phone within 15 minutes of waking up in the morning how many people here actually managed to wait 15 minutes apparently 42% of his admits using email in the bathroom perhaps we do that because we have to spread six hours of email use over the average day one in five people actually admits using email while they're driving ninety-two percent multitask drawing meetings you know we'd like to pretend that we're writing important notes but everybody knows a few emails are coming in and out and what do we do to relax at the end of the day well you're anything like me you get home you sit on the sofa maybe you put something on Netflix but we can't just watch a single show anymore can we we have to switch between that laptop a tablet and a smartphone according to some research 21 times and now most of us are more likely to be fragmented than we are to be focused so what is the solution while some of you may have seen this Financial Times article which came out a couple of months ago about a new generation of San Franciscans who find that LSD in micro doses makes them more relaxed more focused perhaps even more creative so have I got a treat for everyone this evening not really I probably got arrested once and I was interested to read on in this article they interviewed Paul a start-up founder and he feels that him and his team have been less stressed since they've started micro dosing but they couldn't actually be sure about the cause and effects because it might have also been the project management system asana that they started using at the same time and this kind of got me thinking you know if we can't tell the difference between a software-as-a-service and a psychedelic substance well perhaps we're not addressing the root cause of this distraction and fragmentation problem you're staying on track paying attention to the right things at the right time for many of us feels like a battle stay focused on that difficult task or should you check that message it just popped up on your phone are you gonna go to the gym or he gonna stay in bed are you gonna drink tonight maybe not everyone here but I try not to drink on weekdays Monday to Friday but I can always find a reason you know maybe it's a good day maybe it's a bad day maybe it's something to celebrate maybe it's immune to commiserate and often ends up with well just one glass what is going on why does it feel like such a battle to stay on track we know our attention is expressed in rhythms rhythms throughout the day and we can see rhythms of attention that are related to fatigue this is a graph from a fairly recent study that used a classification system to describe how we see this shift away from what we call executive attention networks when we're fatigued we see rhythms of attention based on what we value actually there's some interesting research emerging that suggests that self-control rather than acting like a limited resource actually operates more like cost-benefit decision-making operates more like a valuation process it seems to be associated with something called the dorsal anterior cingulate cortex which you can see highlighted on this little area here now we can actually see and measure these rhythms and I'd almost forgotten I've got this thing on the head so in the great tradition of the Royal Institution I'm going to try and do an experiment and I'm gonna go over here to the corner try and switch over the screens and show you what is going on in my brain so bear with me I'll be back hopefully ok how are we doing yep right there it is this is a three-dimensional rotating model of my brain and this cascade of activity that you can see is actually my brain activity now this is a mobile EEG headset connected to my laptop with Bluetooth is 14 different channels on it and we can use it to measure different brainwave frequencies to understand when attention starts - starts would be really significant where we see the shift away from those executive networks those parts of the brain right at the front of the brain now this particular visualization is more artistic than is scientific to be perfectly honest but I wanted to illustrate the point that actually we do have tools available to us in a way that weren't available before to actually make more of these measurements to actually unlock some of the secrets of the brain to understand what this optimal performance looked like when do we get fatigued what kinds of things are fatiguing is the most today I'm gonna be talking quite a lot about attention a part of the brain at the front of the brain often associated with the frontal cortex so as I'm introducing these ideas keep this rotating model in mind those rhythms of activity those frequencies that are going on for everybody in this room right now because we really just starting to scratch the surface of what's possible I'm gonna disappear again and back to my powerpoint because everybody needs PowerPoint these days then they you can't cope without it right how's that feedback it worked incredible I draw my microphone in the process but that's okay so we can see those rhythms those rhythms of attention but take a step back right now all I realized is headset how fatigued or focused do you feel right now do you have rhythms of work of rest and of play what is really urgent what goals do you value the most what goals are lighting up your dorsal anterior cingulate cortex now most of us find this all too easy for our brains to get hijacked and overwhelmed you know I think these smartphones that we use that can be so much a sports for good potentially often distract and interrupt now I can't give you each one of these headsets but I can give you an opportunity to test your own brain I've got a little activity for us to do everybody ready alert the first thing I want you to do is to remember five words cat Apple truck burger telephone the second thing I'd like you to do is to recite the final five letters of the alphabet backwards and I'm only going to give you five seconds to do it I'd like to do it out loud everybody ready let's go very good now the final task can you calculate 15 times 19 in your head five seconds put your hand up when you've got the answer wow you're a smart group anyone bold enough to share to town exactly now that wasn't the whole purpose in that task how many people can remember those first five words how many people can round all five yeah yeah got it how many people can remember four four was the average I won't ask anybody three now that little game was actually designed to illustrate how cognitive load accumulates maybe you've felt it you see one of the things that interests me as a scientist is how can I measure things and while it'd be great to fit these to everyone's heads and measure their cognitive load and needed to come up with some kind of simple heuristic so this is adapted from a model called the cognitive task load model which describes the load associated with almost any task as the aggregation of three dimensions three dimensions you just experienced in that little game right then the first dimension that you experience that contributes to cognitive load is time pressure if I introduce time pressure into any kind of task I can increase the cognitive load associated with it the second component is complexity I - she was two plus two relative to fifteen times nineteen I imagine for most people two plus two would have come to you a little bit more quickly there's a third component that contributes to cognitive load perhaps it's the most insidious component that component is switching because if I get you to switch tasks and interweave them rather than do them in sequence I'll increase cognitive load I'll also increase the time it takes to complete that task but even if I don't increase the time it takes I'll measurably increase stress so think about the aggregation of those three components to describe cognitive load and then divide it into your in your mind into like all three cognitive gears a low cognitive gear where you might be recovering you might even be task negative you let your mind wander a high cognitive gear where you've got sustained focus where there's minimal distractions and interruptions and then a middle cognitive gear a medium gear where your attention is distributed where you're switching regularly I'd like you to think for a moment about the aggregation of your cognitive load the distribution of your time in those cognitive gears how is your cognitive work distributed in the average day in most of us find that we spend too much time stuck in this cognitive middle gear a middle gear that means we often get caught in pseudo work and switching now I travel quite a lot with my role last year I did about 168 flights across four different continents and wherever I go in the world I'd like to do an experiment it's quite a simple experiment it starts with me buying a takeout coffee and when I buy that coffee I make my order I stand in the queue and I look around me at what people do and there is a global epidemic it seems that nobody anywhere in the world anymore isn't is capable of waiting for more than five seconds without pulling out the smartphone maybe you're doing emails maybe you're looking at social media but actually it's just another form of work for our brains what could have been a moment to look around to interact is sucked up by the smartphone we know as well that in the workplace switching tasks regularly perhaps even unnecessarily increases stress it increases time on task it could also make it harder to switch gears either up to that high gear or down to that low gear if you if you really need to because of something that we call attention residue some of your attention seems to adhere to one task and needs to be peeled off before you apply it to the next one you often it's the people who need to pay attention the most who are the most at risk of distraction multitasking according to some research can cannibalize as much as as 40% of our productive time a typical office worker according to some is interrupted once every 11 minutes apparently we switch our activity once every three minutes sometimes this limitless life that many of us lead may actually be a limiter always being on seems to work quite well for robots which is maybe okay because they're coming for our jobs you may have seen this study apparently in 60% of occupations 30% of the work could already be automated so what about the other 70% that's the thing I'm most interested as someone who is obsessed with human performance well actually I think it is about this augmented approach it's about discovering what humans and machines do best distinctly and integrating it together just as Martha shared at the beginning you can see in this diagram which I'll be happy to share some of the capabilities which humans and machines possess and maybe how we can start to think about they would be distributed but ultimately I think we're looking forward to an age of intelligent automation more specialized roles improved decision-making hopefully increased productivity and efficiency enhanced innovation perhaps but to achieve this we need to make ourselves as human as possible the World Economic Forum published a report a couple of years ago now exploring what kind of capacity capabilities they think are going to be most important in the future of work and they narrowed it down to complex problem-solving critical thinking creativity and coordinating with others as being some of the most important to survive and to thrive but how do we cultivate these capacities day to day and discover a more human way of living and working well we need to find our human rhythms the time of day effects may account for 20% of the variants that we see in cognitive performance in many of us experience a peak it's often best for high gear work where our vigilance is optimal a trough may be best for that low gear work where we really need to focus on recovery and then a rebound now it's interesting during this rebound it seems that actually our inhibition associated with the frontal cortex is reduced and in this this period might actually be very good for insight work and creativity but how often do you think about that natural rhythm during the day and actually try and synchronize your work rest and play with those different phases most of us down now for 80% of people they experience a peak trough and rebound for about 20% they experienced it in the opposite direction you might call yourself an owl if you listen in this category you probably find that you feel quite slow to get started in the morning you probably experience some kind of trophy mid-afternoon like most people but you really get going in the evening whatever the case that trough is associated with many negative outcomes in terms of health in terms of ethics we see some terrible decision-making in many different contexts at that point of day but as I mentioned that rebound period often we waste it but that attentional inhibition might actually be incredibly valuable we might actually be able to use it for enhanced insight but the key is to discover a more human day and actually start to try and live and work and play in line with those rhythms perhaps be more intentional about recharging during that trough really thinking about how you can use medium gear isolate it stop it creeping into everywhere and then also maybe learn how to focus when was the last time that you actually created some space to turn off the devices and work on something in an uninterrupted way I think that many of us need to rediscover how to recover one of the ways that we can do that is actually to start to schedule it in a day real breaks I know in London in the city we're absolutely terrible at this but sheduled your brakes and make it social there's some evidence that suggests that spending time with people that you like spending time with is one of the most restorative things that we could do it used to be called a lunch break but I don't know what happened to those also sure bounced of physical activity seemed to be incredibly helpful to help his recover it could be a walk to lunch with a friend how radical and finally avoiding sue day work perhaps leave the phone in the office at the desk at home and have a period where you engage with the world around you in a different way I don't think we should eliminate middle gear actually it's likely impossible but we could rethink that middle gear maybe start to synchronize the switching activity with the rebound concentrate on choosing menial tasks during that time minimizing it and isolating it but being open to the kind of insights that might emerge during that time perhaps have a notebook next to you to write those things down and finally perhaps most significantly in the future of work to retrain our focus to create times where we guard against distraction and interruption maybe start thinking about working in shorter periods but perhaps most importantly start to work in a way with more precise goals this is a great quote I'm adapted slightly the real danger is not the computers we'll begin to think like people but that people will begin to think like computers a sustainable human high performance isn't really rocket science it's the accumulation of small things small decisions thinking about your time thinking about your energy thinking about your rhythms it's about those small things done consistently well now we've mentioned that we work quite a lot in Formula one and there aren't lots of principles that apply to knowledge workers but there are some and one in particular is a piece of advice that we often give to young drivers who have an incredible amount to pay attention to they're struggling to keep up with us all all this new information they're being bombarded with and when they're getting that that overwhelmed state often we just take them aside and we just encourage them that when they're sitting in that car they just concentrate on the next corner so my question for us all this evening is this what is your next corner well it's your next opportunity to make a good decision a slightly better decision to work in line with your rhythms to maybe unlock some improved well-being and performance whatever the case I think one of the solutions to this is right here at the front of the brain it's your attention your attention which is the key thank you [Applause]

Channel: The Royal Institution

Views: 32,148

Rating: 4.6918139 out of 5

Keywords: Ri, Royal Institution, augmented intelligence, computing, ai, psychology, physiology, performance, wellbeing

Id: JmUFAGgKqjs

Length: 72min 29sec (4349 seconds)

Published: Wed Jun 27 2018