Augmented reveals the stories behind the new era of industrial operations, where technology will restore the agility of frontline workers.
In episode 24 of the podcast, the topic is: Emerging Interfaces for Human Augmentation. Our guest is Pattie Maes, Professor at the MIT Media Lab.
In this conversation, we talk about Augmenting people instead of using or making smart machines, AI summers and AI winters, parallels between AI and expert systems and why we didn't learn our lessons, enabling people to perform better through fluid, interactive, immersive and wearable systems that are easy to use, how lab thinks about developing new form factors, and much more.
After listening to this episode, check out MIT Media Lab as well as Pattie Maes's social profile:
Trond's takeaway: "Augmenting people is far more complex than developing a technology or even experimenting with form factors. Instead, there's a whole process to exploring what humans are all about, discovering opportunities for augmentation and tweaking it in dialogue with users. The Media Lab's approach is work intensive, but when new products make it out of there, they tend to extend a human function as opposed to becoming just a new gadget."
Thanks for listening. If you liked the show, subscribe at Augmentedpodcast.co or in your preferred podcast player, and rate us with five stars. If you liked this episode, you might also like episode 19, Machine Learning in Manufacturing, episode 7, Work of the Future, or episode 13, Get Manufacturing Superpowers.
#24Emerging Interfaces for Human Augmentation_Patte-Maes
[00:00:00] Trond Arne Undheim, host: [00:00:00] Augmented reveals to stories behind a new era of industrial operations, where technology will restore the agility of frontline workers. In episode 24 of the podcast, the topic is emerging interfaces for human augmentation. Our guest is Patty Mays professor at the MIT media lab in this conversation, we talk about augmenting people instead of using, or making smart machines.
[00:00:29] We discuss AI Somers and AI winters, the parallels between AI and expert systems and why we didn't learn our lessons. Enabling people to perform better through fluid, interactive, immersive, and wearable systems that are easy to use and how the lab thinks about developing new form factors and much. Augmented is a podcast for industry leaders and operators hosted by futurist throw Trond Arne Undheim [00:01:00] presented by tulip.co the frontline operations platform and associated with mfg.works the industrial upskilling community launched at the world economic forum. Each episode dives, deep into a contemporary topic of concern across the industry and airs at 9:00 AM us Eastern time, every Wednesday, Augmented the industry 4.0 podcast. Patty, how are you today?
[00:01:27] Pattie Maes: [00:01:27] Hi, doing great. Thank you.
[00:01:29]Trond Arne Undheim, host: [00:01:29] Oh, sure. I'm very excited to have you. And in fact, I just feel like the audience should get to know you. I know a lot of them do because you have become an innovator that, has a stage on Ted and obviously a lot of people at MIT know you. But I wanted to just recognize that, you were one of the early phDs in AI, right? 1987 is not a time. When is that? What we call the [00:02:00] second wave of AI? It's certainly not
[00:02:04] Pattie Maes: [00:02:04] the grandmother of AI.
[00:02:07] Trond Arne Undheim, host: [00:02:07] You're not a recent convert to this topic. That's for sure.
[00:02:11] Pattie Maes: [00:02:11] Yeah. Yes, I actually studied artificial intelligence long before it was such a big deal or the big deal that it is right now. But actually soon after my doing my PhD in AI, I became more and more interested in a related problem. The problem of. Not artificial intelligence, but intelligence augmentation, or how can we make people. More intelligent, more productive support them in making better decisions. So soon after my PhD, I veered more in that direction.
[00:02:48]Trond Arne Undheim, host: [00:02:48] And that's what we will talk about because you have indeed been on the MIT faculty for 30 years, exploring these topics in, various kind of bifurcations. And you have [00:03:00] been the advisor to scores of startup founders also. And of course people might think that goes with the territory at MIT, but the numbers are really still staggering and also the performance of some of those startups, including Tulip.co, which we'll talk about but also many other startups and many other innovation projects that don't, didn't quite make it to startups, but they still created a lot of attention around the world for the promising demos or the things they suggested about what the future of technology might look like.
[00:03:35] So I, I would like first to just recognize that you've achieved, I guess the amazing feat of not just innovating a lot yourself, but you are. You must be an amazing innovation mentor and you certainly have inspired a lot of people that I personally know in AI and in human augmentation and beyond.
[00:03:54] And I wanted, first of all, just to see if I could have you reflect a little bit on your journey, which [00:04:00] I imagine first of all, it's a nice word play from Belgium to Boston.
[00:04:06] Pattie Maes: [00:04:06] Yeah. I came here after my PhD actually, and was of course wanted to be in the place in the world where the most exciting research was going on in my area.
[00:04:17]And so initially I ended up at the AI lab, but I soon after actually adopt or accepted the job at the media lab and what really attracted me there was that the lab is very application driven. We're very interested in really working towards things that can be deployed in the real world that can make a difference in the real world that can be through for-profit startups, but sometimes that is actually other ways by just freely giving away tools and technologies or maybe starting a, not for profits to really disseminate some something and make something accessible to larger groups of people. [00:05:00] So I've always been very attracted to that the practical aspect and trying to make a difference really with the work that we do. And as a result several companies have been created out of my research group.
[00:05:16] Trond Arne Undheim, host: [00:05:16] Was this something you set out to do? Like when you were in Belgium, getting your degrees at the university in Brussels, were you thinking I am going to go to America and become an innovator? Was that in your mind?
[00:05:31] No, I think a lot of that sort of happened accidentally actually. Maybe, one reason, I think why I'm interested in practical applications and real world deployment is that I was never really interested in the technology for the sake of the technology. I'm not one of these people who gets really excited about purely just the technology, the algorithms, [00:06:00] and so on. I want to make my life easier and other people's lives easier and that has always been what motivates me and my work.
[00:06:10] Yeah. And that gets us to intelligence augmentation, because I guess in some sense, the media lab is all about that topic, to some extent and, I wanted to also address the fact that not only are you doing your, the work in your lab, but I think at least for the last few years, you've had the academic responsibility across the lab and you have shepherded the lab, arguably, through one of its more difficult times.
[00:06:35] So surely you have also experienced, innovation and the tricky things that show up with innovation across a plethora of fields. But generally, people at the media lab are hired, because they think about application. What is it that is so different when you, so let's just start with that.
[00:06:53] When you start with the human in mind from the get-go, what is the [00:07:00] difference that makes.
[00:07:02]Pattie Maes: [00:07:02] So I think Indeed's, our philosophy is always to be, like I said, application driven and what that means is that we take a closer look at the ultimate target users and there place or where they live or work and how the technology could make a difference there and could change things there.
[00:07:23] So rather than starting from the technology and trying to maybe optimize some algorithm that does we actually work closely with target users. We really study their lives today to understand what the pain points are, what the opportunities are for technologies to make a difference and support them in being more effective, more productive.
[00:07:49] Trond Arne Undheim, host: [00:07:49] But you have experienced both sort of AI, summers and winters is one of the reasons that AI tends to get into trouble. That it always [00:08:00] is very myopic about The technology focus or AR, or is it a more complicated reason? Why there are these summers and winters?
[00:08:07]Pattie Maes: [00:08:07] I think that is indeed the primary problem.
[00:08:11] So yes, there have been several AI, summers and winters. Probably a lot of your listeners are young enough that they don't realize that there was another hype cycle for AI that happened in the eighties and nineties with the emergence of expert systems, eastward, not based on machine learning and neural network techniques, but instead were typically based on rule-based systems.
[00:08:39] But they were a very sophisticated, they had that typically a lot of knowledge built in about a particular problem. Say making a certain diagnosis or doing some planning or what have you. So these systems were in laboratory settings, we're very impressive and we're often [00:09:00] outperforming experts at doing some scheduling problem or planning problem or diagnosis or recognition problem.
[00:09:07] But what happens when the we're put into the workplace or when people try to integrate them into the real world, was that they basically encountered all sorts of obstacles. One of the obstacles was that people wouldn't necessarily trust the machine the expert system they didn't quite know how to work with it or where to fit it into their workflow.
[00:09:33]They weren't always able to get explanations. For why the machine was making a certain decision. It was very hard to correct or the knowledge of the system and give it new information or to update its information if it wasn't correct. So there wasn't really a lot of transparency, a lot of controllability interpretability and that ultimately was the downfall of expert systems. And yeah, at [00:10:00] that time as well, just like now there were many startups, millions of dollars pumped into all of this. The conferences and exhibits were extremely popular and all of that died down and we entered an AI winter. Where suddenly there was very little interest from the real world businesses in AI.
[00:10:22]Now of course we are in another summer in another hype cycle and I am actually very worried that we are making exactly the same mistakes because most of the AI systems that are being developed are being developed very much not in the context of where they ultimately will be used or not with the collaboration of the people who ultimately will use these tools.
[00:10:51] And so we will encounter exactly the same problems of trust and transparency and controllability and [00:11:00] interpretability in my work, I've always been emphasizing a different approach. And I like to not call it artificial intelligence, but rather maybe augmented human or augmented intelligence or maybe human centric, AI because our approach is one where we start out by studying what people are already doing in a certain work environments, whether that is a manufacturing floor or a doctor in the hospital. And and we actually work together with them or think about how we can support the people that are there to do their work better, to be more effective at their work. And so it's a totally different way of looking at the problem we try to optimize for the person and the technology together to perform better. We don't try to optimize for the algorithm of the [00:12:00] system to become better without thinking about how that system will be integrated into our real lives, so that real-world scenario.
[00:12:10]Trond Arne Undheim, host: [00:12:10] This is super interesting. I want to go into a couple of examples of things that you have done with your students and otherwise in a second, but first, why have we not learned collectively this lesson?
[00:12:20]What is it, is this something you think is happening across the board with technology or is it even just specific to this machine learning AI environment that we is it, are we so tempted by the potential impact of the use cases that would just get carried away? Into the algorithms depth and then forget the user or why haven't people said this is not good enough.
[00:12:46] Pattie Maes: [00:12:46] I think that it is actually a broader problem with development of digital technologies. All of the technologies that we use today, whether it is maybe AI systems or or whether this [00:13:00] social networking services and and so on. They mostly have been designed and built by engineers, by teams that just consists of engineers and not people that come from very different backgrounds, for example more social humanities backgrounds, et cetera.
[00:13:19]One of the reasons that I was very excited to join the media lab as opposed to a computer science department, is that it is very interdisciplinary and we really recognize and try to emphasize that interdisciplinarity is extremely important in innovation, in creating things that ultimately we'll be successful and we'll be able to make a positive difference basically, and a positive impact. So that means involving not just engineers, but also designers people who can really think about [00:14:00] making things fluid, seamless about how it integrates into. Workflow and so on, but also people from humanities backgrounds and social scientists and so on.
[00:14:12] So I think it's important to have that broader perspective to make, or to create technologies that ultimately are desirable and ultimately really improve our lives.
[00:14:26] Trond Arne Undheim, host: [00:14:26] But Patty take me inside of a week in the media lab, because when you describe it this way, it sounds almost so intuitive and simple that I'm wondering why people need to travel to the media lab to learn this. Because if it was as simple to just hire a team with the different skills and it will happen there surely is some other type of magic ingredient. What does a week look like in your lab? How do you draw out the kind of creative energy? It may, maybe it's helpful if you take our novel, Kapoor's alter [00:15:00] ego, which most people know as just that video that went viral and they're like imagining the future of computing with just this device where he's not even speaking, but he's just basically controlling one it would seem the computer with his jaw. Now fantastic video. How does something like this come out of your lab?
[00:15:22] Pattie Maes: [00:15:22] So we are a very open laboratory. So in addition to attracting creative entrepreneurial people and really cultivating a very interdisciplinary team we engage a lot in conversations in discussions.
[00:15:40] With others with the outside worlds, which is actually pretty rare still for people in universities. So for example, we have member companies, we have a consortium of companies that fund the media lab and they pre COVID at least common visits [00:16:00] on a daily basis every day we have at least 10 different companies visiting to see the work, to engage in discussions to give us feedback.
[00:16:10] They don't direct the work, but they can be critical. They can see opportunities for where to take it and so on. And we engage in very Interactive in a very iterative type of style of work where we quickly prototype something like in the case of alter ego, it looked pretty ridiculous. The way it was glued together some cardboard and other things that we could find in the lab, but we create these very early prototypes that are very clunky.
[00:16:43] Don't work very well. But those can, those make certain future more visible they envision what is possible or make it more concrete. And then we invite a lot of feedback from all of these visitors, from all of these [00:17:00] people, with different backgrounds and they see opportunities for, Oh, maybe I would use it this way, or maybe it's really exciting in that application domain or I see this or that problem with the technology. So that's really the technique that we pursue attract a very diverse team of highly creative entrepreneurial people, but from very different backgrounds and engage in a lot of team innovation and two very iterative types of design making prototyping, and then getting feedback from really everyone.
[00:17:40]Not just these companies that come and visit but our own families and of course the targets users of the technologies that we built. So that's yeah the secret sauce, so to speak or the secret how a media lab innovation works.
[00:17:57] Trond Arne Undheim, host: [00:17:57] Take us back maybe to [00:18:00] 2012 or something and in the lab you have two bright people.
[00:18:06] One is Ronnie , who also had a background from the computer science and AI lab at MIT. But then I had already come over to, to study with you. And then you had Natan Lindner who had industry background and had been already head of a Samsung lab in Israel. Now the two of them show up doing their masters, and then ultimately PhDs, but masters, in this context and they start developing something, can you tell me a little bit of what those early days, little early conversations you had with them about what each of them were doing and your reflections on, to what extent some of the work, the early work they did with you? How that transpired into what now? 2014, I believe turned into two lip interfaces and now in 2021 went on the gardener [00:19:00] calendar, essentially as a as a manufacturing execution system and more broadly aspirationally is a frontline operations platform that can transform the way that workers are working at the front lines, augmenting them and really changing manufacturing as we know it today with a kind of a no code system. This was like fast forward, 2012 to 2021. Where were they back then? What was it that you taught them specifically? What were they working on? And how did you work together?
[00:19:32] Pattie Maes: [00:19:32] Yeah. What motivated this work initially was this whole realization in 2012 that we were living in these two parallel worlds.
[00:19:44]We all, and it's still very much the case. We live in the physical world and then there's this whole digital world with information about all the things around this, in the physical world that we are engaged in and so on the [00:20:00] people we're meeting with and so on. And we realized that or we were frustrated.
[00:20:05] So it's really that these two types of experiences were not connected. For example, if I pick up a book I can scan, I can look at the page as the beautiful pictures in the book beater back over to see what people have to say about it but ideally at that moment, I will also have access to the rating on Amazon and what others have said about that book or not, because that's extremely relevant at that moment when I'm considering whether that book may be an interesting book for me to read.
[00:20:40] So we were very interested in creating experiences that are more integrated where our physical lives are more inter integrated with the digital information that exists about everything around us and all of our actions and experiences. So we [00:21:00] experimented with different types of augmented reality systems to bridge that gap and to make the digital information and services available in the physical world. So that's really where the work that Natan and Ronnie did and what led to Tulip.co, where that starts with they were experimenting with building systems that have an integrated camera and projector so that the machine can see what is happening and can project relevant information onto whatever it is looking at so that people can get, for example relevant reviews when they're looking at at the product that they want to buy so we actually developed all sorts of prototypes to illustrate this vision of this integrated augmented reality. We built for example, at that time together with Intel, we built an example of a store that has the two integrates [00:22:00] that has physical products.
[00:22:01]I believe it was cameras and then there was a projector system that would recognize what camera you were looking at or picking up, and it would give you additional information about it. And it would point out the features by actually pointing at the different buttons on the camera and what's so special about them, et cetera.
[00:22:20]We also built a non Augmented desk for a learning context for an educational context. And in all of these cases, we worked with partners. For example, for the education context, to think about how this augmented reality could be used in the context of schools, we worked with Pearson who's the leading developer, of course, books in class books, school books, and so on.
[00:22:48]We then also worked with Steelcase how this augmented reality technology could be used on the manufacturing floor. How could it help people in [00:23:00] real time by giving them feedback about what they were doing, maybe giving them real-time instructions projected onto their workspace, or maybe alerting them that something wasn't done right or a step was forgotten and so on. And that work with Steelcase ultimately then, and with some other sponsors as well GSK, for example which does drug development, all of that led to the spin off to Tulip.co being created as a, yeah a company that can really realize that whole vision of an mentored manufacturing place where you can have real time information provided, but you can also track the whole manufacturing floor in real time and have very detailed data and analytics and intelligence about which steps may cause more errors or where most where which steps [00:24:00] in the process say take a lot of time and so on. So you have this real time insight also into the manufacturing floor that we've never had before.
[00:24:11]Trond Arne Undheim, host: [00:24:11] It's fascinating that they pick this example and that you are explaining it now, because I want to give people the right sense of what it takes to produce a innovation that turns into a commercial true product, because I saw a version of the product you were explaining now in 2014 in the fall when I was at the startup exchange.
[00:24:37] And I was one of the first in there then totulip lab with, seven employees and there, but that demo of something that had a camera and a sensor only this spring turned into what Tulip.co called their vision product, and it's only now coming to market. So here is arguably some of the brightest people working with you a [00:25:00] very illus, a very experienced mentor working from 2012 to a demo in 2014.
[00:25:07]But then they had to take all kinds of other things to market first and only now in 2021. Is this coming out? I find that an incredible timeline and path.
[00:25:21]Pattie Maes: [00:25:21] Yeah it's surprising to me as well. Although I have seen it happen multiple times. We think that technology moves really fast, but then in practice for an invention like this to ultimately make a difference in the real world typically takes 10 years or more. I have had that experience with other technologies that we've invented in the past, actually in an earlier technology that we invented in our lab was recommendation systems. That's say recommend the book to you because you also like these other books or because people who also liked [00:26:00] the books that you buy also about this book that is being recommended to you.
[00:26:05] We invented that technology in 94 when browsers were just available and We were talking a lot to media lab member companies about how exciting this would be and how it would personalize the whole online experience. If you could get these recommendations from other people like you. And and there was excitement among the member companies, but they were at that time saying not sure that people are ultimately gonna feel comfortable giving their credit cards over the internet to buy something. So it just seems very exciting and it's a great vision, but we don't see this happening that was Buster and all the companies that now are bankrupt, maybe because they didn't take this seriously.
[00:26:56] But so because these larger companies [00:27:00] were a little bit skeptical about this whole vision that we were portraying of online commerce and recommendations. And so we started the company ourselves called Firefly in 94 and ultimately sold it to Microsoft actually in 98, but we were just way too far ahead.
[00:27:20] We were too early and Most people weren't ready to buy things online. Most companies weren't ready to partner with us and we actually sold a company in 98 at a time when briefly everybody thought that internet commerce was dead, was not going to take off. A year later, our company, 10 times as much, or worth 10 times as a, as much as what we sold it for.
[00:27:47]So unfortunately we sold it at the wrong time when there was a lot of pessimism about and it's hard to believe that now, especially now during COVID that everybody pretty much buys everything [00:28:00] online. But yeah, back then the 98, that was not at all clear and we were too early. Basically. So in my experience, it always takes at least 10 to 15 years, even for a technology that seems ready to be deployed to ultimately make a difference in the real world.
[00:28:21]Trond Arne Undheim, host: [00:28:21] The digitalization of physical infrastructure like you started with is a different thing though. And even more complicated than the trust to buy something online, which I guess it is vaguely related to you have to trust that something abstract is actually going to have a consequence, but yeah.
[00:28:41]Ronnie and Natan told me that they even, basically slept over in factories and studied these workers for days and weeks on end. And I guess Tulip.co is still studying workers. It's not immediately obvious. What is the contribution on the factory [00:29:00] floor? Is it, it's not as easy as to say we have this fancy digital thing that we're going to give you.
[00:29:06] Why is it so much more complicated?
[00:29:08]Pattie Maes: [00:29:08] Yeah, I think it's always complicated than it is important. That to really understand the context, that actual context of where some technology is going to have to fit in. I remember very well when theRonnie and Natan we're visiting the factories and they would come back with amazing stories about the two are mine's very primitive ways in which all of everything nothing was being done with at that time. Still a lot of use of paper records, for example for collecting information. So it's it was a big gap that had to be a bridge really between sort of the vision that we had of this totally connected manufacturing place.
[00:29:59]With all of [00:30:00] this real-time data real-time instructions and advice, being able to also modify things and edit this whole digital layer or digital support system in real time by the people on the floor and the managers and so on. There was really a big gap from that reality of paper-based systems and a very low tech context to that vision that we had of this smart manufacturing floor.
[00:30:30] Trond Arne Undheim, host: [00:30:30] And how far are we getting with this and how quickly will it go now? Would you say that this has been a decade of exploration and a lot of these things have been sorted out? Or would you say, some quick wins, happened and then some of the slower things, they are just slow with any kind of technology we'll take the time it takes to fully understand how you can contribute.
[00:30:51]I guess I'm asking this in the context of another technology that a lot of people are putting a lot of hope in these days, especially perhaps [00:31:00] during COVID, robotics on the manufacturing floor and that Navy, the merging of AI or machine learning and robotics. How do you see these things?
[00:31:10] How disruptive will any kind of digital device or software system or Augmented system for that should benefit workers. How disruptive can these devices and systems become and have we hit some sort of momentum or is this still going to be kind of case by case basis? And the hype is just not going to be true in this domain?
[00:31:36] Pattie Maes: [00:31:36] I think we have to accept that progress necessarily is slow. I think the potential is there, but in my experience really reaching that potential and learning a lot of, or involves learning a lot of hard lessons along the way. But progress is being made. It's just not as quick [00:32:00] as we would like it to be.
[00:32:02]And I think the same will be true for Sort of this vision of smart manufacturing including the use of robotics, which is even more challenging than because. You have moving parts, which means I think things break down quicker and that there's also more safety constraints and so on as well.
[00:32:22]But yeah I progress will continue to be made and I think it's. Very important for companies to engage with all of these new technologies and to do experiments and to start integrating some of these new technologies in their workplace, or you end up like the blockbuster example that I gave earlier, where they said we'll deal with this later, or when they become more important than they were bankrupt.
[00:32:51]Trond Arne Undheim, host: [00:32:51] It strikes me that you're not going to give me timelines because it depends on so many things. But if you at the future of, cognitive enhancement more generally, or certainly these [00:33:00] immersive and sometimes wearable systems that you have been building for 30 years. You are, you have a, an interesting role because you are of course inspiring a lot of hype just because the products you build are so fascinating and they seem so simple, but you are also combining this with being very careful about the predictions that are surrounding it.
[00:33:20]Yeah. So tell me a little bit about what the future holds for these things. Are we to expect more of these fascinating devices coming on market, or are you exploring a lot more of those in your lab right now? Or, what, where is it at the moment, on the experimental side,
[00:33:39] Pattie Maes: [00:33:39] there's never a shortage of interesting new ideas for us to work on. I always have way too many or more than I have students to work on, but one area that we are exploring in the lab right now is we want to go beyond systems that help people with providing [00:34:00] information, the focus on digital technologies whether it is laptop or watches or smartphones has been primarily communication and also the system giving you information with the work that we talked about so far today the focus was on giving them that information integrates into whatever they are doing so that they don't have to try to juggle between the physical and then the digital information that may be relevant to whatever physical stuff somebody is doing. But we try, we're trying now to go beyond systems that give you information and are interested in looking at how digital devices can help people with issues such as tension motivation memory learning grit, even creativity.
[00:34:54]We think that given that all of us are now forever after. Or we're [00:35:00] cyborgs. We always have technology with us. We have our smartphone never far away from our body. Many of us wear a smart watch as well and so we have this opportunity now to use these systems, to help people with a lot more than just giving them access to information.
[00:35:20] These systems increasingly have sensors integrated. That can sense what the person is doing, where they are, maybe even what their heart rate is. And whether they are maybe a little bit anxious at the moment or not, or maybe the opposite, maybe they're too sleepy. They're not engaged. So increasingly systems will have a better sense like that, of the state of a person, the cognitive state of a person, and will help the person with being in the state that they want to be in. For example, we've been building glasses that have built in [00:36:00] sensors for sensing brainwave activity as well as for sensing eye movements and that pair of glasses.
[00:36:08]It's called T attentive you projects can actually give you feedback about your own attention level. Are you being highly attentive right now or are you being distracted? Are you fatigued and so on and be used that information to help a person to be aware of the fact maybe the, a driver of a truck should be taking a break because they're too fatigued or it can help a person who's listening to a lecture be more attentive because the system can tell them when their attention is waning. We think that this is an exciting new direction to go into really go beyond just giving a person information about whatever job they're doing or whatever. They're working on or are thinking about how we're or [00:37:00] doing, but going beyond that and helping them with those skills that are really important for being successful in life that all of us struggle with and that all of us keep having to work on.
[00:37:16] Trond Arne Undheim, host: [00:37:16] Fascinating. That's fascinating. I wonder, I want to ask you, what is your goal with all of these activities, because you are an innovator, but innovator. So I was motivated good innovator. So it was motivated by something. What is it ultimately that you have been trying to achieve over these years?
[00:37:36] Pattie Maes: [00:37:36] I really want to help people. I did study computer science and artificial intelligence, but my goal is not to create smarter, more capable machines or algorithms. I ultimately want to help people with Machines with AI. I want to enable them to live their best lives [00:38:00] and to grow and learn and ultimately become the person that they would like to be.
[00:38:09] Trond Arne Undheim, host: [00:38:09] So you have a very optimistic view on a future that a lot of people are scared about right now. Some people might be scared about AI they might be scared about what they're seeing around them, how. How do you maintain this very optimistic vision? Is it because you feel like you have agency you're you get clever students come in and work on your ideas.
[00:38:33] I guess I'm just trying to say that usually I would ask people, what is the best way to stay up to date and kind of model what you're doing? And the obvious thing would be, they should try and come and apply and come to your lab. Now, some people will achieve that not very many, right?
[00:38:48] It's a small space. So there's limits. The other advice would be to pay, to get to the media lab and come become a corporate sponsor that seems to be another Avenue. But do you have any other less obvious [00:39:00] ways that people can emanate some of this spirit that I think you, because you're sharing an entire approach to how to understand technology, how to develop technology, but also a vision of what technology should be doing for us.
[00:39:14] So it's almost you have a philosophy. You told me a philosophy with a small P about technology, how should people try to learn more about it, engage with that kind of philosophy?
[00:39:30]Pattie Maes: [00:39:30] I do think it is the role of the media lab to, to be optimist , really, and to see the potential of emerging technologies in improving people's lives. That is really our Unique focus among all university research laboratories. We look at emerging technologies and we try to be positive thinkers or optimistic thinkers in terms of [00:40:00] how those technologies can ultimately empower people to improve their own lives, their communities, and their environment and natural world around them as well.
[00:40:13]We try not to be naive in that quilt at the same time. And we are very much aware that all of the powerful technologies that we work on can be abused and be used in very negative ways as well. But I think that is ultimately not a reason not to engage in, in these endeavors. Basically we, we try to invent the future that we want to live in, or that's really what we are working on. And we try to be inclusive in that process by again, not just involving say the students and researchers in the lab, but really the target communities like [00:41:00] people on the manufacturing floor and how do they want to work with AI and robotics and augmented reality, et cetera.
[00:41:08]So we we basically involved the target users that the companies that are involved in a particular sector and so on as well. And so yeah, I think that there's many opportunities really, for people to be involved. I would also like to say that, especially now with COVID All laboratories have become much more open.
[00:41:31] And for example, lecture series showcases virtual open houses and so on. They, there's no limits to how many people can attend because it's all online anyway these days. So it's actually nice that has opened up the laboratory more and makes it possible for more people to get involved, to be part of conversations, to listen to talks, see demonstrations and so on.
[00:41:59]Trond Arne Undheim, host: [00:41:59] That's [00:42:00] a fascinating, and I think just in closing you, you mentioned this acronym that's typically used in psychological studies, like the weird acronym, Western educated, industrialized, rich, and democratic and it seems to me that is a very specific user group, but it is far from the only one.
[00:42:18]Maybe as in closing, like my last question would be how does one, because other others might be developing technology on other continents or other places, how do you avoid this bias of like jumping into a lane that other people have created that. Is this lane, maybe demos from Western lab cities use cases in highly industrialized factories or whatever it is or created for the New York fifth Avenue, consumer market.
[00:42:46] Those are not the only technologies we should be building. So how do we do it otherwise?
[00:42:53] Pattie Maes: [00:42:53] Yes, I fully agree. And while today I talked about my work and my work is indeed. [00:43:00] Mostly focused on the Western developed world and technologies that might be available here. There's a lot of work happening at the media lab with other communities.
[00:43:12]Both within the United States, less fortunate communities maybe than the ones that my technologies many of my technologies are designed for there's a lot of work, for example, people in Africa use of different technologies. So yeah we try to, we cannot develop technologies for every one.
[00:43:32]But we try to be explicit about who some technologies are designed for. And not assume that they would generally be used usable and we try to work with the targets communities that they are designed for. And definitely we're not exclusively working with or designing technologies for the Western richer world.
[00:43:58]Trond Arne Undheim, host: [00:43:58] Thank you so much that it is, has been [00:44:00] very enlightening. It turns out that advanced technology is complicated and slower, but perhaps more sustainable when it's developed that way and that's an interesting lesson. Thank you so much.
[00:44:13] Pattie Maes: [00:44:13] Thank you. It was pleasure.
[00:44:15] Trond Arne Undheim, host: [00:44:15] You have just listened to episode 24 of the Augmented podcast with hosts Trond Arne Undheim , the topic was emerging interfaces for human augmentation and our guest was Pattie Maes professor at the MIT media lab. In this conversation, we talked about augmenting people instead of using our making smart machines and enabling people to perform better through fluid, interactive, immersive, and wearable systems that are easy to use, developing new form factors and much more.
[00:44:46] My takeaway is that augmenting people is far more complex than developing new technology or even experimenting with form factors. Instead, there's a whole process to exploring what humans are all [00:45:00] about, discovering opportunities for augmentation and tweaking it in dialogue with users, the media lab's approach is work intensive, but when new products make it out of there, they tend to extend a human function as opposed to becoming just a new gadget. Thanks for listening. If you'd like to show subscribe at Augmented podcast.co or in your preferred podcast player and rate us with five stars. If you liked this episode, you might also like episode 19 machine learning and manufacturing episode seven, work of the future or episode 13 gets manufacturing, superpowers. Augmented industrial conversations.
Professor, MIT Media Lab
Pattie Maes is a professor in MIT's Program in Media Arts and Sciences and until recently served as academic head. She runs the Media Lab's Fluid Interfaces research group, which aims to radically reinvent the human-machine experience. Coming from a background in artificial intelligence and human-computer interaction, she is particularly interested in the topic of cognitive enhancement, or how immersive and wearable systems can actively assist people with memory, attention, learning, decision making, communication, and wellbeing.
Maes is the editor of three books, and is an editorial board member and reviewer for numerous professional journals and conferences. She has received several awards: Fast Company named her one of 50 most influential designers (2011); Newsweek picked her as one of the "100 Americans to watch for" in the year 2000; TIME Digital selected her as a member of the “Cyber Elite,” the top 50 technological pioneers of the high-tech world; the World Economic Forum honored her with the title "Global Leader for Tomorrow"; Ars Electronica awarded her the 1995 World Wide Web category prize; and in 2000 she was recognized with the "Lifetime Achievement Award" by the Massachusetts Interactive Media Council. She has also received an honorary doctorate from the Vrije Universiteit Brussel in Belgium, and her 2009 TED talk on "the 6th sense device" is among the most-watched TED talks ever.
In addition to her academic endeavors, Maes has been an active entrepreneur as co-founder of several venture-backed companies, including Firefly Networks (sold to Microsoft), Open Ratings (sold to Dun & Bradstreet) and Tulip Co (privately held). Prior to joining the Media Lab, Maes was a visiting professor and a research scientist at the MIT Artificial Intelligence Lab. She holds a bachelor's degree in computer science and a PhD in artificial intelligence from the Vrije Universiteit Brussel in Belgium.