May 6, 2020

In this introduction to Season 3 we meet our new co-host Lee Hickin.  We also look at what inspired him to get into technology and also look at some of the topical news around pandemic apps, the sharing of data and even play some tic-tac-toe with a WOPR.

Covid Safe app information - https://www.msn.com/en-au/news/techandscience/experts-explain-why-theyre-not-worried-about-covidsafe/ar-BB13oBVC

Paper on contact tracing - https://arxiv.org/pdf/2004.03544.pdf

Ai for Health - https://blogs.microsoft.com/on-the-issues/2020/04/09/ai-for-health-covid-19/

________________________________________

TRANSCRIPT For this episode of The AI in Education Podcast
Series: 3
Episode: 1

This transcript was auto-generated. If you spot any important errors, do feel free to email the podcast hosts for corrections.

Welcome to season 3 of the AI podcast. Um, I promised last time that we were going to mix things up this season. Um, so let's start by introducing my new course, Lee Hickin. Yay. Welcome, Lee.
Hey, thank you, Dan. Welcome. And I'm I'm excited to be here.
Fantastic. I'll let you introduce yourself in a bit, but I suppose people might remember your voice from back in uh episode 19. Uh so your voice might be familiar there and you can tell us all a bit about your kind of history and the things you're interested in and we'll talk about season 3 and the really cool content that we've got coming up. Ray's passed the mic on to you. So he'll be avidly listening in every week I'm sure to make sure it keeps us on track and kind of following us on Twitter and giving us some feedback. So that'll be quite fun as well. So Ray, if you're listening, hi, hope you're well. Tell us about yourself, Lee, from your own general point of view. about your family and where where you kind of come from and what your role is in technology at the minute.
Yeah, no worries, Dan. Thanks, mate. And uh and and thank you, Ray, for for giving me such big shoes to fill. It's uh it's a bit of a challenge to be honest with you, to think about having to fill in your shoes, but I'm really excited to be here. Uh and really excited to get involved in this. I mean, I love to bit like yourself, Dan. I love to story tell and I love to talk about AI. It's a for me, it's, you know, it's just one of those things in our world now that 10 years ago, I probably didn't know that much about it and today it's sort of front of center in my life. So, so the opportunity to be here is super exciting. So, anyway, but yeah, look, a little bit about me and you're right, I did I did episode 19. I think it was interviewed by Ray back then on uh on sort of some of the work I do here at Microsoft. So, I'm here at Microsoft, but much like Dan uh yourself and I am the role I hold here is called the National Technology Officer, which is an interesting role. There aren't many other companies that have a title like national technology officer. Um you know, a lot of companies would have a CTO or chief technology officer which is a kind of similar role but the interesting thing about this role the national technology officer was built into Microsoft's business nearly I it's 20 odd years ago now uh when uh Ray Aussie uh was thinking about this idea of how we should have parts of our business that think about the future and help our governments and our sort of country nation states think about technology and the impact that technology is going to have on social economic and political uh sort of dynamics in the country. So that's where the job came from. I've been in this role now for a couple of years.
Came from Microsoft before, but uh Dan, as you might know, I also took a stint out a worked at Amazon for a little while, you know, great experience. We've seen both sides of the world and but I'm back here at Microsoft because I missed you Dan and I miss you. I did miss you, but not that much. Um but no, I'm back at Microsoft in this role and it it's a super exciting role. Um and I'll talk more a bit about the role, but you know, as you said, from a personal perspective, This podcast was interesting to me because obviously the foundations of it are in education and AI. Um, and I've got family of my own. I've got two kids. I've got a boy who's 13 and a daughter who's turning 10. Um, both very technically savvy. You know, both right now living the very technical life at home through remote remote schooling. Um, and you know, and technology is important to them and it became important for me to think about what does the technology that I work with and and promote and talk about what does it mean to their future, you know, and AI of course front and center to thinking about the future of what our children's lives are going to look like. I'm sure we'll explore that more over the over the season. So yeah, so for me, you know, having two kids has kind of brought a a very personal perspective to to what AI is. Um, and Dan, you know, you and I were talking earlier about devices. I'm not ashamed to say that I have probably got I definitely have more Alexa devices in the house than I do have people. Um, we we use them extensively. I'm a big user of that kind of technology and I love using it and I love exploring ing it. So, you know, for me, technology is home life and work life. But in terms of my role, and sorry, I'm giving you lots of content here. In terms of my role at Microsoft, that NTO role. So, aside from all of that government work where I kind of talk about things like AI and IoT and machine learning and help governments understand how these things operate and what services they could deliver the citizens with them. And that includes education as a sector as well as healthcare. The other role I hold here, Dan, as you know, is I'm the responsible AI chart for Australia. H and that's an interesting role again that's quite unique to Microsoft in that we have an entire established committee from uh from Brad Smith our chief legal counsel in the US all the way through to people like me whose job it is to think about the ethical principle standardized approach to how we deploy and use AI to think about those questions that we should be asking about transparency about accountability about inclusivity and the risks and the rewards associated with AI I and that's a really fulfilling role. I get to talk across the whole business. I get to listen and learn about amazing things that people are building with AI and bring a lens to, you know, to to quote um somewhat of a sort of a an old movie favorite movie of mine, Jurassic Park, just because you can do something, should you do something? And that's a great conversation we have around AI. So, look, I think that's probably more than enough about me.
That's that's brilliant though because I think you know the idea of this podcast as well is to to really bring our own lives and our own experiences into into play here and think about multiple industries and multiple technologies and how that all kind of interrelate us. And you mentioned you mentioned that Jurassic Park element earlier on in episode one. We talked about the favorite AI and I think uh I was talking about Terminator and the the way that actually it was not in terms of really the AI but we were looking at robots at that point and the interesting thing for me was the way that um that technology had actually overtaken society in the future was coming back to sort it all out. Um that was the kind of interesting juosition I was kind of going for from there. But now you've mentioned Jurassic Park and I have to ask you about your favorite AI and what like influenced you to learn more about technology and AI.
Yeah. Uh yeah, of course there's a lot a lot of influences from that time of my life in the 80s. I'm a I'm a child of the 70s, but I grew up in the 80s. So yeah, all that kind of stuff's really important. Okay. You know, you're the first person that's ever provided me with a positive spin on the Terminator story. So, you love the idea, the fact that the AI came back from the future to fix things that we'd all messed up. I like it. That's a great way to think about it. I'm I don't know if I can be that positive. Um, oh, look, yeah, deep 80s sci-fi fan. And funny enough, when I talk a lot about AI to customers and to the market, we get into this conversation of of of 80s sci-fi or at least the 80s interpretation of that has really shaped a lot of actually how we think about AI today. You know, our perception of what AI is generally speaking in the population is driven by things like the Terminator and Skynet. You know, how many times have you seen the Skynet meme as the way to describe something about AI? It's it's very common.
It's very it is. And so, um your question about what is my favorite one. Uh again, we're back into war and fighting. So, I'm a I'm a huge fan of the the movie War Games. I I 1983 was the movie.
Yeah,
I love that movie. Um, I was sort of a little bit obsessed, I guess, with the idea because at the time I was uh I was 12 years old. I was a nerd on my computer tapping away all the time. That's what I did with my time. And so the idea that you could you could have that impact at the time, you know, with computers was sort of world expanding for me. So if anyone remembers for you for those of you who want to go do the digging in war games, there was a computer called the Whopper, War Operations Plan and Response Machine, which was the worst example on video of a computer. If you ever remember, it was a sort of green box of flashlights.
Um, that's right.
But that was an example of an AI that would play out these scenarios for World War II. And obviously, you know, the positive element of that, of course, is it worked out that at some point there is no right answer. There is no way to war at the end. Yeah, that's right.
That's right. So, you got that idea that, you know, something as complex as rich as what do they call it? Thermonuclear war is a comparative to tic-tac-toe in that there's a futility to it. So, I guess that for me was the, you know, and uh look, I was so obsessed with that movie, I ended up actually naming my son after the child in the movie. It's called the the doctor, the guy who invented the Whopper. His son was called Joshua. My son's
I thought you were going to call him Whopper then. Yeah. Joshua.
Yeah. Like like a Big Mac Whopper. Tom, but um but yeah, look, movies is a big influence and and you know, I think it's interesting. I don't know if you've discussed this in previous podcasts. Haven't listened to all of them yet myself. Um but there's a saw a whole bunch of really interesting influences that kind of got me interested in it. Um, you know, big reader of Isaac Azimov sci-fi and Arthur C. Clark stuff, you know, kind of the content around
uh the laws of robotics and and all that kind of stuff. And it's fascinating stuff for me. You know, I grew up in that era of that time of space exploration and technology just kind of burgeoning. Um, I'm also a big uh Douglas Adams fan. Uh, so if you any Douglas Adams fans out there, you would remember obviously Marvin the paranoid android, the uh the the dress AI, but also the uh the serious cybernetic corpse genuine people personalities. All the doors on all the ships have a personality and AI personality.
Yeah. Yeah. You remember? Yeah. And it's
absolutely
it was just a funny take on it. I thought that was uh fascinating. So look, you there's so many things, but you know, the 80s negativity I think is is an interesting myth we have to dispel, but you know, it's a good place to start in terms of people understanding kind of what what is AI cuz it It gives at least you a context for it. You see a robot talking, you see a thing interacting,
you get a sense of that what AI is and we can explore more of that I guess through the uh through this interview.
So currently I know this is going to date this episode but I think so thanks for giving us that insight into yourself Lee and and it's really I love the way that that when we think about AI everybody's got different perceptions around it but it does end up going back to 80s films. It must be it must be the showing age of everybody that's joining the podcast but But also like in the current climate, uh I suppose one thing that that's worth noting for this episode for the for the people tuning in is obviously we going through this epidemic at the minute and and I know we've had quite a lot of conversations internally and you know in the media around all the applications that people are using now around uh the COVID safe apps and things like that. And obviously we don't want to go down any rabbit hole today of of what what's good, what's bad. There are but it is interesting how the different technology come companies and the different governments uh addressing this issue and and just from a NTO point of view with your role at the minute are there things you can kind of share with everybody today around what your thoughts are around this and I don't mean about any particular type of technology but
what about how can we de deconstruct that the technology behind some of it because some's AI and some may not be or yeah just just interested on your take
yeah no absolutely and you're right I mean the role because of the role and the way I get to work with governments and help agencies and others. I I see and hear a lot of good stuff and it's I think that the thing that's for me Dan has been the most um I guess heartwarming uh you know kind of gives you some faith in in the way we think as a civilization is just the the massive surge of everybody wanting to help and everybody wanting to contribute collectively uh without any kind of look for commercial outcome generally speaking that's the way things have been I mean the early example when you think about the co safe app uh the co safe app there is the work that Google and and Apple did, you know, they came together and said, "Okay, how do we collectively help build and deliver this construct around contact tracing, which is the,
you know, kind of the idea that we're trying to build with these tools." Um, so I think that's that's kind of a critical thing is that, you know, and and we Microsoft like a lot of other companies just went to government, to health agencies, to those that are dealing with the challenges and said, "What can we do?" You know, what do we need to be able to do to help you either sustain your business through like we're using now through you know technology like teams or what do we need to help you do to better understand the problem you're trying to solve in the health sector or whatever. So that that for me is kind of the big the big piece. Um but you asked about the co safe app and it's interesting because obviously that is both something that is on everyone's minds I suspect uh you know in terms of should I shouldn't I what does it mean what and there's a lot of talk of risks and all these kinds of things as you said I don't think we need to unpack all of that that's a you know probably a rabbit hole that we want to get into.
I think it's important to understand And when we look at it, why we are being asked to do it? What what's the need for the app itself? And and that fundamentally is this idea that is widely accepted health communities around the world that contact tracing is the most effective way to establish and understand the patterns of the virus itself. You know, so the contact tracing idea is that if we come into close contact with each other and we uh you know, one of us contracts the disease that we can quickly and easily identify the spread pattern and and uh and eliminate it.
Um, but what we don't have here, and this is the thing I think a lot of people be concerned with, is that the app, you know, it's it's gathering data about me. It's learning where I go and what I do. It's going to capture all this information about the people I see and the places I go. Look, that's patently not true. And there's been plenty of deconstruction of the app on Twitter and in other places that explain and the app itself, you know, it's built off an open source bit of code. You can go look at the open source code. It doesn't do those things. It really is just designed to do things. So, I think the key thing here, given we're on an AI podcast, is to say that the co safe app is not an AI app. You know, it's not there to artificially learn about what you are doing to build some profile of you as a citizen. It's kind of there to understand what you do and at the time when you choose to allow it to um you know, keep you safe. But it's I I think it's a really, you know, good example of one of those scenarios where something we could never have done before. Technology and data coming together to create a safe environment in a situation that is, you know, unprecedented in this pandemic at least in our lifetimes the situation I mean I'm it's sort of it's a food for thought if you think about a 100 years ago in the time of the Spanish flu
uh and the and and the imp impact that had globally around the planet but we would had no technology no ability to track tracing no ability to weigh these things yet still millions of people died and it's a terrible tragedy
but I wonder what we would have had
with the technology and the ability to just quickly and easily identify um you know people at risk so yeah oh yeah it's an interesting It it is and I think the other the other interesting thing for me and I know there's a there's a paranoia about privacy and things but I think that's also quite a telling sign from the communities as well because I think people are now more concerned about what data governments have on them or third parties have on them. You know over the last um couple of years we've had the things that are happening with with elections and in the US and things like that in the UK through Brexit and there's a lot of things. So people like were really concerned about that privacy element and, you know, not even using it. People are thinking, you know, about the lockdown laws and if they'd be lifted and how are they going to manage, how governments now going to manage society and really going down like a a a kind of dark and deep path. But it is, I suppose, encouraging, especially from um my my side when I'm working with schools a lot and working educational establishments who really care about privacy a lot. It's good to see people having that level of conversation even though sometimes people uh can get quite paranoid for right reasons or wrong reasons like we did the same thing for the national health record in Australia didn't we uh like people still thinking about that so you know I think it's a good good the people are pushing back and saying well what what is happening with these applications because
it shows that the society generally is becoming more aware of how people use our data and things.
Yeah definitely yeah well and that awareness yes absolutely it's growing and and you know in some ways it's arguably good to have a healthy sense of skepticism of of any, you know, kind of of anything in your life just to to validate and check. And I think, you know, AI has been it's delivered a lot of really good things. There's no question about it. AI in healthcare, AI and humanitarian work, AI and just helping us really get to the heart of some problem, some big humanitarian problems around the world. Super valuable.
The way that the people working together is quite interesting as well, isn't it? Because I remember in one of the podcasts just jumped into my mind there. We were talking about the Bushfire app and the the app in New South Wales was different to the app in Victoria and I was in the snowy mountains right on the border. The Victorian app was telling me to get out get out of dodge and the New South Wales app said you're fine, you know, which app do I use? How does this work? What are your thoughts on that kind of, you know, the the way different states in Australia or probably countries might be starting to share data and things?
Yeah, it's a it's a really good point and obviously you know, something like the Bushfire app, which is, as you say, state based is a is a really dangerous example of how, you know, the great thing about having a federated set of states as we do here and, you know, similar to the US, is that you get these boundaries of of autonomy and ownership and and, you know, and New South Wales looks after New South Wales people and Queensland the same and so on and and it's good from a
kind of distribution of of function and value and and and being building government services that are applicable to individuals, but as you say, in situations where we have a a national or in this case a global uh scenario going on. Um you can start to see the cracks. You can start to see where data not being shared or integration of systems. You know, it's kind of like I think there's always a I don't know if it's a um an old wives tale or an urban myth of the idea of that. Yeah. When the when the British started building train track when the British were over in the US and they were building train tracks, they had two different gauges, you know, the sort of the the two different size of train tracks and nobody ever thought that the fact that when they neat that they might actually not work because it's two different trains if not even if it's a madeup story. It's a good example of this idea that we can have data in New South Wales about uh you know bushfires or in this case um you know healthcare record data
and if the data is different in Queensland even if we shared the data it just wouldn't work together you know there's that challenge of integration and integration costs money creates complexity and often obuscates some of the data value.
Yeah and when we're looking at education we're going to see we we're going to have between schools as well because every single school or every single university kind of you know are are different and they do have their unique cultural values but underlying that there's a data schema that are common between most and you know that data sometimes isn't surfaced you know internally uh within the school and certainly often not shared uh among systems. So this is a really interesting kind of take on it. So when we're looking at the um the the co are there any other things you've noticed in terms the trends that have that have come up. Have you seen any AI being used or any types of technology that you think are kind of jumping out?
Yeah, know it is. Look, there's there's lots of things jumping out, you know, in AI in particular. I think one of the things that we may have sort of we touched on this earlier is this if you think about what AI is and you know that we should definitely do an episode about what is AI and really break apart some of these moving parts because it is a phrase we just use, you know, without thinking that we talk about AI and as you and I both know there's detail behind that behind that term. But, you know, we're seeing AI being used because it's good to spot patterns in data at large scale. That's a a very basic construct. If you think about it, AI can look at a huge set of data that you or I could never get our head around and see the most minute pattern occurring over time or over a dep, you know, depending on the type of data and then provide that insight to help, you know, a human clinician in this case or someone else identify an atrisisk patient or identify a hotspot flare up or identify a predetermined set of conditions that are likely to cause the risk of a a resurgence of the the co 19 uh virus. So that's an interesting dilemma. You know, in this scenario, we're all sort of talking about the co safe app and privacy and the concerns and the asurances of our own civil liberties, but then at the same time in the health care sector uh and and largely in governments nationally, internationally, we're actually seeing this opening up and they're saying, "Look, we've got data about our patients and our citizens and our cases of co let's share that with other countries other places to better understand so they can be better prepared or we can better understand our own and it's a really good example of you know when you have data at scale and you have the access to the AI tools to sift through that data for the patterns you know you don't know what you're looking for then you start to really enable this sort of global data sharing idea and we're seeing that come together now we're seeing in the US sharing governments and they're sharing government and Microsoft has been working with uh the White House and with University of Washington and others to create a platform for open data set sharing as as have many other vendors and places and we're sharing the data not not just plain data but the labeled data so people can use it very quickly to sort of make decisions. So I think that's one thing I have noticed is that you know as we know data underpins the value of AI and data being shared is more broadly happening because we understand now the value of of how quickly we can move to things like co 19 situation with the right data and the right tools to to do it you know and the issue of privacy is not doesn't go away in fact we should maybe put it in the show notes but we've we've written an article Microsoft um chief scientist a guy called Eric Horvitz has contributed to a paper on the protocols of privacy around contact tracing so thinking about how do we do contact tracing but also maintain the privilege the privacy of of individuals as we do it so we're always still thinking about the the need to have the tools, but also the right to maintain that. Um,
I've also seen quite a few universities and schools and things bring up chat bots. I've seen quite a lot of that myself, which is quite interesting. I remember I did a LinkedIn post on how to create a chatbot because it's it's quite easy. But, um, yeah, lots of universities and people doing remote learning have been creating chat bots to help interact with it or interact with their university or interact with their lecturers. Quite interesting use of chat bots and through governments, right?
Chat bots are a funny thing. I mean, yeah, we again a whole I think you you've already done the conversation the episode we're coming now
but um you know we think about chat bots in the in the co world you know people are using chat bots to just get answers to questions quickly people are largely worried scared concerned and don't know all the information and we don't have the capacity for the beer person on every phone to answer every single person's phone call but a chatbot can deal with that level of of uh you know influx of calls but also and this is the thing I again I think about what AI can do chat bots can communicate to you in whatever language you need to speak. They can address their answers in the pace that you want to respond to them at. They can deal with people that have limited communication capabilities or speak different languages or just are not comfortable communicating with a person perhaps around something sensitive like their health. So you yeah there's there's a there's a lot of good in that as well as just the kind of the you know robotic voice kind of mindset people might have around chat bots.
Yeah, absolutely. And and this I read something about the research accelerator program and the open data set you mentioned there earlier. on is that is that something that specific is that a global thing or is that uh US only or is it a way that everybody should
it's global and yeah we should definitely uh put some notes in this but sort of again you know thinking about the things that I love doing about my job and and why I why I came back to Microsoft to take this role on is because we just constantly look for ways to contribute back um into society into the world we so you you'd be familiar I mean you may have talked about it before the AI for good program
which is you know bordering program of investment for finding ways that AI can help improve the lives of citizens around the world and everybody you know regardless of country and and and everything else.
So one of the things we've done with the COVID research accelerator is under the AI for health program we've kind of built a specific stream of that invested a ton of money into that program such that we can really target where COVID research vaccine research uh COVID research into um into hotspots into identifying communities at risk uh or into simply just helping governments better understand the scale of the problem. Uh we put a bunch of money and process into that under the AI for health program uh to help organizations get access to Microsoft people resources and technology and we've coupled it with the uh an opening up of the Azure data sets program which is a tool we have in the cloud as you'd know for people to host open data. So data sets that are available for everybody NASA data in there we've got New York cat taxi data in there whole range of stuff. We've opened up that for COVID data to be shared globally. We've made it free for customers to upload that up to 10 terabytes of data.
We'll help you clean the data, meta label it, and tag it and then put it up in there so it can be shared uh globally.
That's that's fantastic, isn't it? I think, you know, the things that people are putting in place at the minute and you know, we were talking before this podcast about our own kids and remote learning and whether that's really going to change education. And I think there's another question in there about how this is going to actually um change the way we address these global pandemics as well, isn't it? Because um after SARS and and some of those other epidemics there there was a lot of uh issues around how those actually you know obviously in terms of the diseases themselves this isn't an epidemic podcast but you know the diseases were fundamentally different so they are much more different to transmit but actually when you're thinking about what technology companies and generally society have done and governments have done to try to deal with this. Um I think it's going to put us in a better situation for pandemic too or as as well I was speaking to one school leader this week and and you know they were relaxing the laws in the state that he was in which is in South Australia and they were starting to go back a little bit to normal and he said actually I'm actually doubling down efforts because he said this is going to come back in a couple of months and he wants to make sure that you know they've learned from the mistakes they've made trying to rush things through and make sure that they prepared for pandemic two which might happen in three months time, five months time, five years time. So, okay. Yeah, I know. Yeah. So, he's waiting for this, you know, just imagine him storing away some sandwiches and some tins of tuna. Um but um yeah, thanks again, Lee, for for um supporting us with this podcast going forward.
It's great to hear from you. Yeah. And and when we look into the season ahead, I suppose just a bit of a flavor. Season one um we we looked at some of the general implications of AI and very much on the basics of what these things were. AI for evil was one of our fun podcast we did at that point and really tried to explore some of the ethical issues around AI. Season two we started to uncover well now we know a little bit about AI then what are some people doing in industries such as government in the police in healthcare very generically and then also some of our partners about actual customers and data scientists are actually doing in this space um to really drive that and and we were sitting down yesterday uh to think about what we're going to drive for season 3 and we were really thinking about lifting the lid on the technology and looking it from some of the emerging areas as well around say quantum and aiming to unpack some of the some of the technologies under under the hood I suppose and really try to make this as educational this season as educational as possible as well as bring in some experts in from the feel. What are your thoughts on on the season ahead, Lee?
Yeah, look, it's honestly, mate, that's why I signed up for it. I'm just really excited by uh, you know, what you've done already that the stories you told over the other episodes are fantastic and it really is that journey of kind of walk you walk, crawl, walk, run in terms of your AI skills. And we're kind of getting to that point where, you know, we're going to start running and that's great because we're going to get into the details. I'm I'm so looking forward to talking about quantum entanglement, um, stuff like that. Super exciting stuff. But no, I think I think it's looking good and and yeah, we want to get a little deeper, I guess, in some of the technology, but still keeping it to the point where, you know, it makes sense and we can all kind of relate to it. Uh, I'm looking forward to I think that and I think we can maybe even think about talking to some people that are that are really doing AI in the real world today, practical AI that's being used in business and and bring some of that to the journey as well. So,
fantastic.
Looking forward to it. Looking forward to it.
Thanks. Thanks again, Lee. And see you in the next episode.