Language selection


Innovate on Demand, Episode 5: Survival of the Most Adaptable (DDN2-P05)


Is there a lack of innovation in the Government of Canada? Fact or fiction? On this episode, guest speaker Sinan Baltacioglu discusses how ones willingness to adapt to new and emerging technologies can contribute to success and innovation within a department.

Duration: 00:34:30
Published: November 8, 2019
Type: Podcast

Now playing

Innovate on Demand, Episode 5: Survival of the Most Adaptable

Transcript | All episodes
Listen on: Apple | Google | Spotify | Stitcher


Transcript: Innovate on Demand, Episode 5: Survival of the Most Adaptable

Todd: I'm Todd Lyons.

Natalie: I'm Natalie Crandall.

Valeria: I'm Valeria Sosa.

Sinan: And I'm Sinan Baltacioglu.

Todd: And this is the Innovate On Demand podcast.

One of the myths of the public service is that governments lack innovation, efficiency and effectiveness when compared to the private sector. Our guest this episode worked for a mega-corporation before joining the public service and experienced first-hand how any organization can be bogged down as it scales up in size and complexity. He believes that success is tied to a willingness to adapt. So, what has his experience been like since joining the public sector? Let's find out.

Natalie: Thanks for joining us today, Sinan. Do you want to tell us a little bit about yourself and where you work?

Sinan: Absolutely. It's a pleasure to join you guys today on this podcast. Thanks for the opportunity. So, I'm Sinan. I'm from the Digital Academy. Specifically, I'm on the Digital Innovation Services team. You can kind of think of us as the R&D arm of the [Canada] School [of Public Service].

Natalie: Very interesting. What kind of things are you guys up to right now on that team?

Sinan: Like all good geeks do, we are deep into experimentation and doing minimum viable products and doing proof of concepts and really trying to push the envelope when it comes to the art of the possible. So, what does that mean? It means we're experimenting with technologies like machine learning. We're trying to do sentiment analysis on free text so that we can get quantitative information instead of just qualitative. We're pushing the envelope on what it means to do quick builds. So, you give us a napkin of a couple of ideas that you have, and then you walk away and 4 days later, we come to you and say, hey, it's in the cloud, here's your application, check it out. We're working on agile processes, trying to change what it means to be a developer in the Government of Canada.

Valeria: How long have you been with the team?

Sinan: I'm actually a pretty new transplant into the public service. I started in January, so I've only been here about 6 or so months and I'm loving it. Prior to that I was in the private sector, so getting the opportunity to use your time and talent to help Canadians and the public service and really do right by your country is an opportunity you just can't get anywhere else. I've been loving my time in the public service and looking forward to continuing.

Valeria: And what were you doing in the private sector, if you don't mind my asking.

Sinan: A mega-corporation is where I spent the last 7 years of my life. I was in startups. I was doing diabetes management software—think of it like Facebook for diabetes—a lot of reverse engineering for metres, and so on. So, anything "geekery-wise" you can think that we've done it in the past.

Valeria: And what did you study?

Sinan: I'm actually a software engineer by training. I went to the University of Waterloo from 2001 to 2007. Did that and basically there's no breaks. So, it's like 4 months of school, 4 months of co-op, 4 months of school… And I think that kind of prepares you for life in government, where it's just go, go, go, go, go all the time.

Valeria: Uh huh.

And how did you find the Digital Academy? Or did they find you?

Sinan: Funny you should mention that… I was sitting on my couch one day going through my endless Netflix playlist, which is impossible to get to the bottom of (I think that's an NP-Hard problem. Someone on the line can check this for us). But I was on Reddit and I find this article about how they’re creating the Digital Academy. And I was kind of paying attention to the government because I saw the Canadian Digital Services coming around, and [I thought], interesting signals are happening in gov. And then the moment I saw what the mandate of the School is, which is to upskill 300,000 public servants, I'm thinking, Well, this is exactly it. Because I've been developing since I was 14. I built my first corporate webpage in 1997 and I'm only like, 35, 36. I've spent nearly half of my life doing tech stuff. But it's kind of neat, because now it's getting to the stage where anybody can make amazing things happen super quickly. So, being able to be at this juncture in history and say, Okay, now I can snap together things like AI systems and sentiment analysis systems and just kind of go with it. It's super wild.

Valeria: For the record, not anyone can just snap things together! [laughs]

Sinan: That's my hope, though! I want to see a future where you don't need a PhD in order to get involved with these new technologies like machine learning, for example. In my first week of working for the public service, my boss, Sean Kibbee, who's the director of the Digital Innovation Service… Sean's a great guy; we love him to pieces… he looks at me says, Okay, we've got this idea. Let's make a chat bot. And I say, Okay, what do you want it to do? He [says], I don't know, it should do this and this and this. Figure it out. Let's try something. Stand it up in 3 or 4 days. Okay. Off we go into the "diss pit" and we start banging away on code. Now, I don't have a PhD, I'm not a machine learning expert, by any chance. What I do is, I've got an ability to use metaphors and analogies to help people understand what it is from a human perspective. So, I went into the system, I'm thinking, Well, I don't have time to gather a couple of million data points for English and French and try to understand the language. And well, how is it going to recognize my speech? And then once it does that, how's it going to understand my intent? And then I realized, you don't actually have to do any of that stuff yourself anymore. HTML5, it's got speech recognition and transliteration. So, I can say something to my phone and then it turns it into text. Well, that's step one. But now, how do we get the intent? There's a system called I throw my text at Wit and it returns something to me, which says, I think I'm 95% confident that you wanted to search this site. And then what I do is I take that content and the intelligent part is around what you do with it. And that's really where everybody is going to kind of stumble with AI and machine learning. Because if you use the wrong data, or you use too much of a specific type of data, or you didn't massage it the right way first, the answers that it gives you will be right, according to the machine, but may be wrong according to us.

What do I mean? So, let's say you wanted to identify dogs versus. wolves. You give the machine a million pictures of dogs and a million pictures of wolves and you say, Okay, machine, I've let you do this unsupervised learning, figure it out. Here's a new picture. What is this? And you give it a picture of a wolf in a living room. And then the machine tells you that it's a dog. This giant thing is a dog. And you ask yourself, why? Why did it get it wrong? I gave you a million pictures of dogs. I gave you a million pictures of wolves. You should have figured this out. And then when you start poking it, you figure out you didn't actually build a dog/wolf checker. You've built something that checks to see if the background is a snowy Norwegian forest. Because if it is, it's a wolf. A machine will just simply try to give you the answer that it thinks will answer its own question. For example, if you say, Machine, give me world peace. A completely viable solution is to eliminate all the humans, because that's world peace. No one's fighting. It may not be optimal for us! But that's around defining the context and giving the machines a little bit more about the human side, which is really going to be the challenge of the future as we move forward into this stuff.

Valeria: Interesting. Talk to me about something that you're working on now and how it's going to apply to the public service or public servants or how they can use it.

Sinan: Sure. So one of the neat projects we're working on right now is actually an interesting little thing that we did where the Deputy Minister came to us. He said, We want you to take a look at the way we collect information from our learners,- from the people who attend our events, from the people who attend our courses. Think about how it's done today and instead of doing an incremental improvement on the system, we want you guys to take a quantum leap into the future. Think blue skies. Go out there, build something, put it in front of real human beings and come back to us with real data about what works and try something. Just don't sit in a room and have a meeting to have a meeting to talk about the requirements to define the requirements. You guys are smart. We trust you. We hired you for a reason. Go out and do it.

Valeria: Did you [shed] a little tear when they said that?

Sinan: Oh, absolutely. I like to think of myself as almost like a fire-and-forget missile. You give me good instructions and I will give you what you need quickly. And then we can iterate on it and make the product what you want. I don't like spending years on ideas. Because by the time you get to the implementation phase, the entire environment might have changed. And your idea, while great a year ago, isn't great anymore.

So, what we started to do was, [say] Okay, what's the hard part in surveying people? Some people are going to throw their hands up and say, Well, every time we give a new survey to IT, it takes this many weeks to get done. And then we have to go back and forth and fix it, blah, blah, blah. And I say, I actually don't think the technology is the hard part. It's an annoying part. It's frustrating. It shouldn't be. But it's not the hard part. It's routine. It's rote. The hard part is asking the right questions. The hard part is talking to the internal team saying, Okay, what are we going to survey? Do we need this to go through Ethics? Do we need to have certain people come in and chime in? And so everybody usually builds a survey in Word first and we email it around, we talk about it, we correct the words. Okay, we shouldn't ask it this way. This is exclusionary language, we should do it like this. This is masculine coded. And you move on. And that's the hard part. So I said, What if that was the challenge? Step one is to redefine the way the business process works. Why do I need developers to build a form? Forms are generic. Let's build a language to describe forms. So, imagine if you were to build a full web-enabled mobile application that works on every device, multimodal, blah, blah, blah, what would you need to do? I think it's as simple as saying: Q, colon [and] this is my question. Pick one. It was good. It wasn't good. What if that was all you needed to write to build an application? You don't need to be a developer. You know what questions you want to ask. Q: is pretty common. So, that's what we did. We built a generator that takes this language of surveys, generic, that you can literally email to people or Slack to people. And with a copy-paste, we can drop it in, and in 10 seconds, generate a form which is instantly out there in the cloud. And then the major bonus is, building dashboards is really annoying too. Because, you build a form, and then the dashboard person builds a dashboard, then you change the form and then the dashboard breaks. Well, let's not build dashboards: let's generate them. So, now we use the data that we get and we automatically build these full custom dashboards which show you the information on the day. So, imagine this: you've got 50 people in a room, they open their phones, and they scan a QR code. And within 3 seconds the form is on their phone. They submit the answers and on a screen somewhere or on your tablet as you're walking around, you see the answers coming in live—bing, bing, bing. Ninety percent liked but 10% hated it. Now, before they even leave the conference room, I can put my hand up and say, If anybody really hated this, I'd like to talk to you, if you'd like to talk to us. Think about the opportunity you get now. Now you have live, real, valuable feedback. So, that's what we're trying to do. That's the E-Valhalla project. So eval, E-Valhalla. We're dorks! Our API is called Regis-Thor. [laughter]

Natalie: Sinan, I have a question for you. You mentioned that you worked in startups before. I'm really interested in the concept of innovating in the federal government, and what are the opportunities and some of the obstacles. I'd like to have your take on what the differences are in the private sector and in the government, and where maybe we have room for improvement or opportunities to seize here.

Sinan: Absolutely. So the first thing I'm going to say is, it's kind of a mistake [to say] that the private sector has it all right. The moment you increase in size to the point [of a] mega-corporation, the issues and challenges that we have in the public sector are pretty much the same. Our projects are complicated. The people we're working with have a lot of deadlines and pressure on top of them. Our mission changes. We get people coming and going. There's a bunch of shift and drift. But, forgetting all that for a second, the startup mentality is really kind of a philosophy around: if I don't do it now, if I don't do it well, I'm not going to eat. I used to joke, when I was working in the startup, [that] I'm a mercenary and I work so that my cat can have cat food. That is my win-condition. That's my end goal. Because if I went home and I didn't bring the cat food, despite the fact that she doesn't have opposable thumbs, I would wake up the next day not here. Definitely shuffled loose the mortal coil.

Natalie: I have a cat, I know.

Sinan: Yes. I think some of the interesting things that we can transfer from startups in the private sector into the public service is [to] give people faith and trust in their abilities, [and] when you're hiring them, you hire them based on their capacity to learn not just what they know right now. And then the real challenge—I'm seeing a lot of resistance in the public service so far, but I think that's just because we're going through this transformational phase right now—is, you can't give chefs rusty, dull knives. So, if your machine is even 4 or 5 years old, with the acceleration and the pace of change, it's not just 4 years old, it feels like it's 40 years old, because now the system that you need needs 4 gigabytes, but your entire machine has like 2, maybe 4 if you're lucky, and it starts falling down. If you need to crunch data and you're working at Environment Canada and … your machine falls down and crashes every time you're crunching this data, how effective can you be in your job? That's time of Canadian citizens that's going to be wasted. That's taxpayer money that disappears, if you think about it. I love developing on computers which pretend to be tablets. Nice and light and small. I can touch the screen. I've got full freedom to do what I want. I can install things on my machine that I need. I can work in Python. I can work in R. I can work with these open-source tools. For example, at our department, GitHub isn't blocked, but there are certain public departments where you can't access GitHub. If you're in one of those departments and you can't access GitHub, I want you to talk to your leadership right now and tell them to change today. Because that's like someone saying, Oh, I've got work to do. I'm not going to pay attention to these horseless carriages. What is this stuff? My horse gets me where I need to go. [laughter] And then, you know, I don't see horses in New York right now, aside from the few novelty ones. The game is changing. We need to catch up. It's no longer survival of the fittest, it's survival of the most adaptable. For the public service to survive, we have to adapt and adaptation is going to be scary, the change management process is going to be hard, all of us are going to have to upskill in things that we used to say, I'll never have to know this. This isn't interesting to me. I think every human in the public service needs to get a "little d development" in them, where they can say, Okay, I understand this data a little bit better. I've got more data literacy under my belt. And it's a common language where we can start. There's going to be no end state where we're finished. We are now entering continuous learning and if you stop learning, you're done. When I started building "modern web pages" in 2007 or 2009, I needed to know like 4 or 5 tools back then. Now in 2019, I need to know 33 tools. It's insane now. There's a Cambrian explosion of technology and it's only going to accelerate with the changes in AI, robotics, gene editing, mass social communication, the fact that you can have a conversation with a couple billion human beings at once. There are now more mobile phones than there are human beings, if my numbers are correct. That's wild, if you think about it, because the human brain has 86 billion neurons. What happens when we get 86 billion devices out there? Well, that's the complexity of the human brain and that's when scary stuff starts emerging out of it.

Valeria: Let me ask you, what do you "nerd out" on after work hours? What's your passion?

Sinan: I've got a couple of passions. When I'm not on a development station, working on software and code, my other development station is a big ceramic, charcoal barbecue. And you laugh…

Valeria: I laugh because I'm from Argentina. I appreciate someone who appreciates barbecue.

Sinan: I like to joke that because I'm a software guy, when the power goes out, I have no skills. Because you can't program without electricity. And no matter how much [tippity-tapping] you make on your keyboard, maybe you can make some interesting music, but I'd rather use a drum set. Oh, well. So the advantage here is I like starting fires with flint and steel, throwing chunks of carbonized trees into a pit and then spending the next 8 or 9 hours smoking half of a cow. That's what I love doing. [laughs]

Valeria: Oh my God! [laughter] You're speaking my language! We call it "Asado."

Sinan: Asado? Fantastic! I'm still learning the whole barbecue arm. I got a new one just recently. My folks got together [and said] Hey, here you go. This is a BBQ; make it work. So I've got these 8 dinosaur Flintstone ribs in this barbecue. I've got it going at 235 [degrees]. And I did some research into "the stall," which is that at 150 degrees, the meat just kind of holds for a couple hours. I thought, How do we fix this? This is what engineers do. How do I fix the BBQ process? Well, this is an engineering problem with inputs and outputs. I thought, Well, okay, humidity… the longer it sits in a humid environment, the longer the stall lasts. So, let's put apple juice and cold filtered coffee in my water pan, but only 1.5 litres, so by the 3-hour mark, it's completely gone, so it dries out in there. Then we've got 3 hours of drying action, so you get the smoke absorption for about 4 [hours], the drying action on the rub for 2… only crack it open twice to keep the heat up. And lo and behold, I got this amazing rib that came out with a nice pink ring around the outside, this homemade, delicious barbecue sauce with wasabi and Saigon cinnamon… This is what I mean. When I'm not designing programs or building software, I'm attempting to do the same with my palate.

Valeria: Your brain is fascinating! [laughter] By the way, we're going invite you over for a barbecue.

So, let's go back to AI. What do you think about AI in the government context and what's going on?

Sinan: AI is an interesting thing, because it’s turned into one of those blanket terms now. It's almost like rock and roll. You can ask anybody, Do you like rock and roll? And the majority of people will probably tell you yes. However, I like to listen to Swedish melodic death metal. That's also rock and roll, kind of. So you've got to be really cautious when you say, Oh, is this artificial intelligence? Okay, what kind do we have? What we're dealing with now, in today's day and age is usually "artificial narrow intelligence" [or Weak AI], which means you give AI a picture of a dog [and] you ask, Is it a dog? Okay, here's a bunch of text: tell me if the sentiment is positive or negative. Identify this, categorize that, classify this, give me an expert opinion on that. It's very much problem:solution. But when you say artificial intelligence, you have to remember what we are comparing it against: natural intelligence. Well, what's natural intelligence? Well, it's about 1.5 kilos, about 86 billion neurons, about a trillion connections and the darn thing runs on 12 to 20 watts. Yet, it's a dual-core massively parallel processor that creates sense, reason and imagination. And it's between your ears. And it constantly evolves and updates its own software, it can sustain significant blunt force trauma and still work. You can cut the bandwidth connection between the 2 cores and this thing still works. It's one of the most amazing processors out there, and the darn thing runs on bananas and salt, potassium and sodium. I mean, that's what we're comparing AI against. So, we're not there. The next stage of AI is the artificial general intelligence. And this is more what people think about when they say AI, because they think, well, I'm going to ask [it a question], and it knows. But humans are very bad when it comes to anthropomorphic things because we want to believe in magic. We want to believe that the Wizard of Oz is real. We don't want to believe that there's some dude or gal behind the curtain who's banging out code and saying, Look, it's magic. It does all this. And really, it's a homunculus. It fools you [into thinking that it's intelligent]. It's the Turing test concept. If you ask it enough questions, eventually you see the holes. General intelligence is when AI actually starts understanding context. And it starts being able to have metacognition: thinking about the way it thinks, grabbing disciplines from other zones and adding them in. And at that stage, you also get a bunch of the other disciplines in machine learning, artificial intelligence, that whole arm, which is machine perception, movement, social intelligence, and all these different Lego pieces start snapping together. And it's the collision of these things that make it amazing. We had cameras, we had phones, we had email systems, we had faxes, music players, video cameras. It wasn't until we [combined] them all into one form-factor device, which was super small and light and easy to use, did we see amazing changes start happening. We're going to see the same thing with AI. Once you get into general [intelligence], then you get to the spooky stuff—artificial super intelligence. Some people argue we're never going to get there, which might be the case. I think human beings are super-duper creative and solving one problem solves many other problems and creates a whole bunch of other ones. But as long as we survive our technological adolescence, and we don't be dumb and kill ourselves off and have a nuclear winter or a genetic [catastrophe] with CRISPR. There are so many different end-game concepts which are floating around right now. It's really exciting. So, you asked, "What do I geek out about?" I geek out about end-of-the-world scenarios. So, you'll find me reading up on Ragnarök or Ecclesiastes theology.

Valeria: Okay, so I'm going to rein it back in and just ask you about the government context and AI. Give me a few points of advice that you would give the government with respect to AI and what's going on now. Or lessons learned or clarifying "This is what you need to know."

Sinan: The first thing I'm going to say is, do not be fooled by the Wizard of Oz. The private sector will come to us and they're really good at checking RFPs and RFQs and checking the boxes, [and promising that their product or service will meet our requirements]. And if you don't have the PhD, or the deep understanding of machine learning, it's very easy to say, Yeah, I trust you. You're a trustworthy person; a human being. You've told me it does this thing. But it's nuanced. For example, you say, I want you to build a system for dogs. Show me dog parks. That's reasonably easy [with] geolocation. Show me dogs in parks. We only added one word, but now I need a PhD in computer vision and who knows what else to solve this problem. In government, [the lesson] is make sure you don't get tricked by the flashy side of AI. Because it's very easy to build a 10% solution that makes you feel like you have 90, but to get to the 90 takes a lot. So, first and foremost, be skeptical. We're still in the Atari days of all of this technology. Number 2, use AI, machine learning, intelligent analysis and predictive analysis, to help inform decisions, but not make decisions. Because again, the important part here is a computer or a machine will optimize [problems] based on its parameters. Understanding human context, what matters to us, is a little different. For example, you go to a restaurant and they say, Oh, your table is free. I know that this means there is no one sitting at my table, and I can go there and sit down. But the machine [may interpret this as] it costs no money. Lack of context. The last thing I'd [suggest] is [to] genuinely experiment. Start trying these things out. Don't be afraid of failure: fail forward, fail intelligently, fail smart. Give [employees] the runway and freedom to actually inform policy through experience. I want there to be someone who sat down with the system and built something and tried it with a couple of real humans and said, This was good, this wasn't good, and this is [what we should do differently]. If you can inform policy through experiences, it's a lot easier. Watching Google demonstrate Google Home is one thing, but having it in your kitchen and you scream at it to add wasabi to your shopping list and it keeps adding Wausau [Wisconsin] beets! It's [frustrating]. What are you doing? Understand me! I'm not just singling out Google. It happens with [Amazon] Alexa. It's all of them. Any speech recognition [system] will have this challenge because we're all a little different. We all speak a little differently. If you use colloquial terms, your AI is going to be [trying to figure out], What did they say?

Valeria: You should hear me try and get Spanish music to play with my Google Home. I have to say it with an English accent!

Natalie: Hey Siri, send me the directions to 123 "Princi-pally" Street, because apparently a lot of these are not bilingual software.

Sinan: Yeah. "John Dee Ark."

Natalie: I have something I'm really curious about, Sinan. I've been really interested in robotic process automation recently. And I'm actually starting to think about how robotic process automation and AI actually work together, particularly in the government context. I feel like we have huge opportunities in the government around our processes and repeatable things. And I think the core group of public servants who work in administration as generalists who are supporting most of the operations of the federal government are going to see some unbelievable changes coming in the near future.

Sinan: I completely agree. 110%. I like to say, let humans do human things, let machines do machine things. What machines are really good at is repeating something perfectly, every time, quickly. Let's say that you need to check underneath trucks to see if the wheels are okay. Maybe a human being going around [a truck] is going to take a couple of minutes to do this. Well, a quick little drone or a robot goes around and gets answers for us one way. But it's not even [limited to] physical robots, like industrial style, like the Canadarm-type robots, which have become even more amazing. (In fact, I highly recommend you go to YouTube and check out this robotic arm dancing with a human dancer. It's a very interesting thing where you see the human training the robot, and then the human starts to falter a bit and then the robot takes over and supports the human. So, that's neat).

But for us, the automation isn't just for robotics. Let's say you're in HR. And let's say you're just drowning in the firehose… digital innovation services just sent you 45 requests to do and then this other department did the same. You just can't get out from under it. Well, if you start using automation to figure out, What do I do today? What do I not need to do manually myself? What can I reliably put into this automation framework? And then once it's automated, the freedom in my time gives me the ability to do something new. So, let's say that I don't actually need to go through all my HR documentation to validate whether or not the person who's applying is a graduating student or eligible for a co-op. Could you imagine if you got 1000 applications and you can send back 10 requests to the hiring managers to say, Oh, yeah, by the way, the person you're trying to hire graduates right now so they're not actually eligible for a co-op. You need to do a casual. I really think the automation around our processes, automation around anything that we physically do… I know we're thinking about using drones and stuff like that for environmental checking, using the topology to give us better data. You can automate that. And then just speaking to your point about how AI chips in. You can build a robot, and we've built robots before that assemble things; we've had them for a long time, but they haven't been AI-enabled. What it would do is go up one metre, grab something, go down one meter, twist something, and put it on shelf. Imagine for a second that the thing that it was grabbing wasn't exactly where it was supposed to be. Well, instead it grabs Frank. Now it puts Frank on the ground. Now it rotates Frank and throws him in the cabinet. It doesn't know! But an AI system would know how much the thing it's lifting weighs. And then it lifts it. Oh, wait a second. That weighs 120 pounds. That's Frank! Don't throw Frank around! If it notices, it can adjust. So the ability [of robots] to use data and sense the environment and their context to inform where they are in space and time and then how to change their responses…. One of the most amazing things to watch is an elephant, for example. It can be so gentle with its trunk and you just watch it moving a branch without snapping it. This is an animal that's a thousand pounds and could just cause devastation, if it wanted to. Seeing a robot do something very similar is an uncanny experience. Because you know, it has more power than we ever could. But here is this almost-human thing going on when it's just gently doing these things. I think a lot of that comes down to how do we use robotics? How do we use AI to improve what we do today? But the first thing we have to do is model it effectively. Because if we don't actually know what we do, or what's important to us, we're going to build solutions that don't actually fix anything.

Valeria: That's very interesting. I've been recently talking about people's perception of time and being able to improve work environment. And for me, I think people have a distortion of their perception of time and it inhibits them. [They think], I don't have time to make things better. To me, it's not an excuse because you have to work towards the future as you're working towards the now. Because as people discover what work is going to be left if the machines take over, if you can sort of decipher what automated things you can get rid of, then you can allow yourself the time to do that extra thinking for improvement.

Natalie: It's amazing. I mean, the art of the possible for me in my head right now is the synapses firing on the robotic process automation and the AI marriage. And I'm thinking about, you know, Estonia, where you have these tiny little digital bots who actually have their very own credentials, and they go into the 10 or 12 different systems that need to be accessed so that a citizen can perform one task, or one process. So the citizen just says, Yeah, I need to do this. And then there's a digital AI robotically automated process—intelligent agents—that have their own set of credentials, their own username, their own password, and they have a proxy right to [access] your account and your data. And they go in and they'll apply for whatever government service that you're accessing as a citizen. I look at that and I think about the applications in the government. Just thinking about our current context of the Phoenix situation, and the HR and the leave and everything. Imagine, I could have a little bot who could go and actually start mining, with my credentials, all of those systems and compile my file for me or something like that. I think it's just mind blowing.

Sinan: My dream is I want to have a bot for the School where you say, Hey, I'm interested in learning "blah." And then what it does is it farms the information that we have, it uses input from people who've taken courses or found stuff on the net, and then it tells you, Hey, if you want to get good at Python, for example? Start here, then do this, then go to this website, go do that, then go to YouTube and watch this. I want it to help shepherd me to that knowledge and when we start talking about automation and the usage of AI and the social intelligence around what this stuff means and extracting context… The future is really bright—if and only if we work together and make sure that humanity is part of the batter before it even goes in the oven. That's how we win.

Valeria: I feel like every team needs a "you."

Natalie: Thank you so much, Sinan.

Sinan: Oh, it's been a pleasure, guys. Thank you so much for having me on the show.

Valeria: Thank you. It was fantastic. And you're welcome back any time—and over for a barbecue!

Todd: You've been listening to Innovate On Demand, brought to you by the Canada School of Public Service. Our music is by Grapes. I'm Todd Lyons, producer of this series. Thank you for listening.


Todd Lyons

Valeria Sosa
Project Manager, Engagement and Outreach

Natalie Crandall
Project Lead, Human Resources Business Intelligence
Innovation and Policy Services

Sinan Baltacioglu
Senior Technical Advisor
Digital Innovation Services, Digital Academy

Tell us what you think

Share your comments on this episode by using our feedback form.

Related links

Date modified: