Overcoming technology challenges of AI and Robotics
- My name is Peter-Anthony Pappas and I'll be moderating today's panel entitled Overcoming Technology Challenges of Artificial Intelligence and Robotics. The panel we'll be touching on a plethora of AI related topics which should be exciting, especially considering the passion of our two panelists. And regarding our panel, I'm honored to introduce my two esteemed guests, Elnaz Sarraf and Nisha Talagala. Elnaz is the CEO and Founder of ROYBI and the award winning ROYBI Robot. ROYBI is an investor backed EdTech company that raised $4.2 million and its seed round focusing on early childhood education and self guided learning through AI.
ROYBI Robot the robot is the world's first AI powered smart toy to teach children language and STEM skills. It's been named one of Time Magazine's best inventions in education placed on the 2019 CNBC Upstart 100 list as one of the world's most promising startups and included in Fast Company's 2019 world changing ideas. Before starting ROYBI, Elnaz co-founded and led a consumer electronics/IoT company iBaby, which she serves as the company's president. Elnaz is also a board member of the Consumer Technology Association, Small Business Council and is a member of the Forbes Technology Council.
Her honors include being named as a NASDAQ Entrepreneurial Center Milestone maker, being named the Woman of Influence through Silicon Valley Business Journal and being named Entrepreneur of the Year in Silicon Valley. She's also spoken at several conferences such as the Mobile World Congress, the ASU, GSV Summit, the Consumer Technology Association and more. Growing up as a woman in Iran, Elnaz witnessed limited opportunities which led her on our journey in the U.S. to become an entrepreneur and to create a technology that would empower children by providing universal access to personalized learning and an education that prepares them for a better future. Thank you so much for joining our panel, Elnaz.
- Thank you so much Peter for having me. - Absolutely. Nisha Talagala is the CEO and founder of AIClub, an EdTech company which brings AI literacy to K-12 students, professionals and other individuals worldwide. Nisha has significant experience in introducing technologies such as AI to new learners from students to professionals. Previously, Nisha co-founded ParallelM which pioneered the MLOps practice of managing machine learning and production for enterprises in which later was acquired by DataRobot. She also started the first MLOps conference and registered the respective trademark which is now owned by DataRobot.
I always try and give trademarks a plug whenever possible, so there you go trademarks. Nisha is a recognized leader in the operational machine learning space having founded the U-S-E-N-I-X Operational ML Conference which is the first industry/academic conference on production AI/ML. Nisha was previously a Fellow at SanDisk and Fellow/Lead Architect at Fusion-io, NVM Software Lead at Intel and CTO of Gear6. Nisha has more than 20 years of expertise in enterprise software development, distributed systems, technology strategy, and product leadership. Nisha earned her PhD at UC Berkeley in Computer Science.
She holds 73 patents, over 25 referred research publications. Is a frequent speaker at industry and academic events. Is a contributing writer to Forbes and other publications. Thank you so much for joining our panel, Nisha. - Thank you very much.
- Wonderful to have you. So too small fun fact before we begin, prior to meeting each other to discuss this panel, even though Elnaz and Nisha are extremely well known in the field of EdTech, they didn't know each other nor did they realize that they're practically neighbors living in the same city. So, you know, we're bringing people together here at the PTO, perhaps future addition to our slogan. So it's kinda neat, small fun fact there. As for myself, I currently serve as Dennis mentioned as the supervisory patent examiner at the USPTO.
Previously, I served as a special advisor to the Under Secretary of Commerce for Intellectual Property and Director of USPTO where I advise the director and other USPTO executives on IP and operational matters such as the impact of AI in IP policy and the use of AI tools to aid in examination of patent and trademark applications. I currently serve on the USPTO office of the Under Secretary AI Working Group in the USPTO Patents AI Point of Contact Team. The latter of which oversaw the development of an AI based prototype search system at the PTO. I also serve on the White House National Science and Technology Council, NITRD, AI research and Development Interagency Working Group.
So jumble of words there. And on the Department of Commerce, ICSP AI Standards Coordination Working Group. Elnaz, turning over to you, is there anything that you'd like to add before we start? - I think what'd you said was amazing. And I have to say, I am truly impressed by everybody's achievement here and once again, thank you for connecting both of us, three of us actually together.
And I'm really excited to share my story today. - That's awesome, Nisha? - I think I would just echo everything that Elnaz said. You know, really looking forward to the discussion. I think it should be amazing. - Agreed, so now that everyone know enough perhaps to be dangerous, I suppose let's move on to the questions then.
Elnaz, if you don't mind, I'll start with you. You know, in our prior chat, you shared your personal story of an inventor, would you mind sharing a bit more about that journey and its origin, please? - Absolutely and thank you for asking you about that because I always believe our stories are very important. You know, I moved to the U.S. about 15 years ago from Iran and I don't know if you guys now but life in Iran is quite different that what it is in the U.S. and you're talking about 15, 20 years ago.
You know, back in Iran, I always wanted to be different, do something different. I always wanted to be an entrepreneur. I always love to build things, you know.
I studied art back in Iran and then I always wanted to bring my art to life. So I switched my major to software engineering and I went to one of the top third universities in Iran and was able to actually learn a lot about technology. And both art and technology really allowed me to have the opportunity and understanding how to build products. But of course, I never had the opportunity to reach my dreams back in Iran because of so many limitations for women.
I hear things are a little bit better than 20 years ago. But you know, when I came to the U.S., I absolutely grasped the opportunity because I really felt as a proud citizen now I have to say, it's the land of hope and opportunities. And I saw the opportunities, I started even going door to door, giving flyers to people, trying to bring businesses, understand the culture of the U.S. and that's how my journey started.
And I was able to co-found one of my first company which we have baby monitors and all of that experience being around parents, learning about their needs, led me to found and create ROYBI Robot which I will be talking about it more today. - That's awesome, thank you so much. And Nisha, you previously shared with me that you were awarded your first patent 20 years ago, you said and currently have 73 issued patents, I think that's right.
You know, so clearly you're still a computer scientist with over a decade of industry experience. The entrepreneurial spark that motivated you to leave that all behind, so to speak, and branch out on your own. - Thank you, great question. So, I started sort of life classically as a techie, right? I was a technical person, I loved math, I loved coding and things like that and I got my PhD at UC Berkeley because I was in computer science because I was really into the topic. And I wasn't quite sure really what I wanted to do except that I wanted to do technical stuff. And I worked in a number of different companies and one thing I think I've always done sort of in my career is I looked for something that would help me grow and would challenge me and frequently chose that over anything that felt safe.
So after working at Sun Microsystems for a number of years, the next company I chose to join only had 12 people and it was quite a mental shift to go from a company of 30,000 people to a company of 12. And I did that because I knew that they couldn't really stop me from doing whatever I was proved I was able to do. You know, in a large company you have a job and you really cause a lot of trouble if you do things beyond your job, you know? Whereas in a small company, there's just too much work to be done and not enough people, so I kinda knew that I would only be limited by what I was able to do. So I think that's really when I got my first feel about entrepreneurship and I joined Gear6 as an architect and I actually became their CTO within about two years because I worked really hard and I coded a lot of the product and stuff like that. I love to innovate, I love to create new things and personally I like to do it in a small environment where you build things from the ground up and there's a tremendous sense of joy and pride and accomplishment that comes out of that and so, this is the second startup that I founded and that's sort of how I found that this was possibly the right fit for me.
All those-- - You know, in education and it's so personal because for me actually, so when I was growing up, especially back in Iran I was really fortunate that my parents paid a lot of attention to my education. So I had the opportunity to learn different skills like painting, music, sports and even I was studying English language. So when I actually was growing up, I could utilize my skills so gradually work on my dreams but also when I moved to the U.S., I was immediately able
to get into the workforce, get into the university and do so many different things that I couldn't really do if I didn't have enough education when I was a child. So with all that passion, I always felt that early childhood education is a very important and critical area for children. And when I was talking to my co-founder, he shared exactly same vision. He has three kids, he always talks about all the difficulties are experiencing with school, with programs, especially with homework, you know? So we decided to start actually doing something about this category. And when you look at the early childhood education, you don't see good tools and products to help children at this critical time.
So that's really how we got started. And the reason we got started more into the conversational and language learning aspects of AI and to use AI to help children to focus on their interests and abilities was that we realized there's a huge gap. A lot of children have for example, one in 12 kids in the U.S. alone, they have some sort of speech disorder but there are no good tools to really help them. They only have to go to therapy sessions, a lot of them they don't like it.
So many, many of these reasons actually, it's not just one reason really led us to create ROYBI and be able to have an impact in the education space. - Yeah, so I think that is really an awesome story. So from my kind of area, so my parents are teachers actually.
So my father is a mathematics teacher. My mom was a physics teacher. So I've always sort of been around education and oddly enough, I paid my way through college by working as a math tutor. So I would tutor other kids in calculus and stuff like that.
And so, I've always sort of been around education, it's always been something I've enjoyed doing and things like that and then how I got sort of into AI literacy is that I've been in the, first of all, I've been a fan of AI for a long time and I did it sort of in my spare time even when I was working on other topics. And I have a daughter, she's 13 now and even while I was in computer science, I tried very hard to get her to be excited about what I was working on. And before I worked on AI, it was simply impossible because I was working on like things like operating systems, really hardcore stuff, deep down in the bowels of complicated systems, completely unrelated to her.
Then I realized once I started working in AI, that she could relate to the things that I was working on, ooh, you know? Companies make recommendations, you know? When I go to Netflix, it makes a recommendation. Oh, that's AI. It's related to me.
And I tried very hard as Elnaz pointed out, you know, I tried in the AI literacy area, many years ago, I tried very hard to teach her AI. I looked at all the available tools and they were just too complicated for her. And she got bored and this and that and I ended up doing a lot of side work to try to create things that were not boring for her. And that process is what led me to drive AIClub the way it is because I realized that it was actually possible for them to grok this stuff.
It's all around them. It is actually safer for them. It's better for our future if they understand these technologies but you cannot present it to them as a mathematical textbook. You know, because frankly I can't deal with it that way, much less than anyone else. And so, we were able to sort of, our AI Literacy Program and why we call it AI literacy is that you really want kids and the next generation who are AI natives by nature.
I mean, they have never been in a world without AI around them. Now we can all remember times without phones, times without the internet but they have no such thing. AI has been in their lives since the day they were born. And so, they need to be literate and making them literate does not mean they have to learn partial differential equations or crazy math, it simply means that they need to understand how it works, what they can do with it, what can they build, how they can connect it to their imaginations, things like that. And so, that's how I got into kind of the AI literacy and that's how AIClub came about. And my daughter now loves AI, she's built a ton of stuff and she's worked with friends and things like that and so, it's now very much something that she's conversant with along with thousands of other kids.
So that's sort of how I got to AI literacy. - Yeah and that's great. And as you said, a lot of it too is just bringing the awareness to people that AI is in their life so much.
So many people think of AI as this autonomous car driving or Skynet, let's say, but it's been around for quite a while and it's a lot of the mundane things for lack of a better phrase that it is there. And I think that was a great point, it's just sort of bringing the awareness to that. It's in your life already, you don't have to go to AI, AI is already-- - AI has come to you. - Yes. - That's right. And so, that's really neat. Okay, great, thank you.
So Elnaz, I know you mentioned while we were on, of course, I'm gonna ask you about ROYBI Robot. So that's where we're at now. Could you tell us a bit more about ROYBI Robot, its past, present and where you see its future, please. - Of course, thank you. You know, when we started the idea of ROYBI it was in 2017.
I still remember, I had a napkin in front of me and I was drawing some sketches and I was telling everybody, this is gonna work, this is the future and everybody was like, "Okay, let's do something about it." And I also remember when we were pitching to so many investors, we constantly said, AI is going to be the future of kids learning. The future of education.
And we would get many interesting comments from the community, from the investors. But we actually officially started a company in 2019. From 2017 to 2019, we spent a lot of time on research and understanding what we can really do to help children.
I remember our very first idea was more offline, very similar to the Robot Jibo. It was more about entertainment, having a lot of content, with the display to just for children to have fun but we realized that's really not about the future and about the children. So we finally designed ROYBI as you can see it here, we spent about four to five months on the design itself which we also of course, have a patent on this. We wanted to make sure it is small.
It is not heavy for children to hold it. And since we are focusing on ages between two to eight, it's really important that the new coffee is trustworthy and it looks like a companion friend. And what he does is of course, it's an AI powered. It's conversational, we focus on language learning, communication skills and also some STEM activities.
In terms of AI, what is so important about our technology is we also hold the patent, utility patent on this which is very exciting, is our Edge technology. We are actually literally the only company in the world to have Edge technology which is on-device voice recognition for children. You know, it's a very, very complicated area. It took us about five years to build the algorithm, train the engine of course. We've done over 150,000 children and of course, compliant data.
And it is something that we've been building because when you're in children's space, it is very important to pay attention to their privacy. So everything gets done on the device, no delays, and it is very innovative. And for children, they play with it on a daily basis. You know, they interact with the device.
They talk with it. ROYBI can understand the answers, it can respond back to them. We also have an app for educators and parents to see their progress and all the reports of course, are AI generated.
Another interesting part is of course, we expose children to sophisticated technology and AI from very early age. And where we see ROYBI going? Of course, we are working on more programs, more languages. Now we have English and Chinese. We are adding Spanish, German, French, and Japanese next year which is going to be very, very exciting. And of course, we are also looking to collaborate with government entities, with schools in order to get ROYBI to as many children as possible, to have them have access to quality and fun education.
- No, that's great. And you sort of beat me to a later question about EdTech but we're gonna get to that later because that's an interesting point and I think for our audience they might find that actually where emergent technology has its own emerging technology. So do you have an interesting story that you could share with us that maybe others might not know related to ROYBI? - Oh, of course, you know some of our exciting moments are the times that our customers actually reach out to us and they tell us how much ROYBI is helping their kids. One parent they reached out to us, they said their child is an autism spectrum. And she doesn't really talk with anybody much and she actually uses ROYBI, she tells ROYBI all about her day. She sees ROYBI as a friend and those things are so motivational for us.
Another parent reached out to us and said their child has speech delay and it has been helpful, of course we don't have metrics yet but all of these amazing stories that we hear, it's really helps us to build even better programs. And now we are also thinking about establishing some programs for kids with autism spectrum. It's really crazy, you know 20 years ago, there were one out of about 20,000 kids that they were diagnosed.
Now it's one out of 54. So these numbers are just alarming and we just want to make a difference and when we hear these stories, just get us moving forward. - That's great, I didn't even know about that. Is a medical space an area that you all are, the medical community rather, something that you're gonna look to partner with perhaps? More so, because of that? - Perhaps because we actually had a couple of parents told us that their kids actually take ROYBI with them to therapy sessions.
And I think that is something that we are going to explore in the future and then see how it goes. - Great, thank you. Nisha, could you tell a bit about AIClub which I see in your background there. And also on Pyxeda, I'm not sure if I pronounced that right, AI, which I believe is somehow tied to AIClub but I'd love for you to elaborate on that.
And again, same thing, it's past, present and where you see its future. Thank you. - Sure. - Yeah, so just to clarify that, so Pyxeda is our corporate entity and AIClub is really our offering.
So they are one in the same but we usually go by AIClub because that's really what everybody interacts with us about. So I thought that maybe I would start with the fun story first and tell you guys a bit of fun stories, that way it'll kind of give you some context. So I guess several, then I'll in various different ways. So one of the things that I'm recently most excited about is you know, one thing we've been able to do is to reach kids around the world.
And I'm originally from Sri Lanka by the way and so we had a girl from Sri Lanka, her name is Anudi. She built an AI with stuff that she had learned from us and it was a chat bot for the lonely. Bot called COLBY which is essentially an AI powered chat bot, she noticed during COVID lockdown everyone around her was alone, stuff like that.
And one of the really nice things about this is that she made the finalists at a international competition called Technovation Girls where she was one of six children out of 1700. And the beauty of it is she was the first from her country and it has energized like an entire collection of kids her age in Sri Lanka just because they watched her and when they heard the final presentations, I think like her entire school joined in to watch her present. So it was just this wonderfully motivational thing for her, that was sort of one of them. And another one I think that was really fun was a feedback kinda you know, I also get really awesome feedback from parents and sometimes they're just very interesting and one of the feedbacks that I got was from a child, he's about 12 years old and the mom said, "You know, we went to a party and he was talking "with a friend of ours who was a Google engineer "and they were talking about AI and he was keeping up. "And the two of them had a conversation I didn't understand "a word they were saying (laughing) "but they seem to understand each other "and it was sort of really impressed." And so, one of the things that I really kind of interesting is how, I'm always really impressed by how much kids can do.
You know, if you just let their creativity run loose. So going back to your question of how did we get started, where do we go. So we've been teaching kids now for a little over a year and a half and we started by actually building a technology platform that we call Navigator which is essentially an overlay tool over industry tools. So powerful industry tools like AWS or Google Cloud platform. And what happens is that those tools are ultimately the tools that adults use.
You know, however, they're clearly not accessible to children because they're too hard. So what we did is we bridged the gap, we built a tool where any kid can access it through their web browser, they can build an AI in five minutes and they do, you know. Every single kid in our classes regardless of their age build an AI in the first 30 minutes of their class. And they talk to it and they interact with it. But the nice thing is that this is not a toy tool.
It is actually a tool that is simply a front-end layer over professional tools and as they get older, we peel the front end away slowly. And by the time they become high schoolers, they are literally dealing with professional grade tools. And so, this has materialized. And so, what we find is, kids seven to nine to 10, all front-end overlay, they don't see the backend tools. Middle school kids, see a little bit of the backend tool.
High school kids, they're in the backend tools. So it just basically kind of really nicely bridges the gaps so that there's no like cliff that they fall off of like I've outgrown the toy tool, I don't know how to use the grown up tool and this and that. And so, that's the technology fundamentally that we built that power sort of everything that we do.
And then on top of it, we've created curriculums, you know, lots of exciting use cases. And because of the automation, every student can build their own custom projects. So literally, like nobody does the same thing. So by day two, they are all doing their own thing and this is amazingly empowering for children because what they're building is not, you know, it's something that is entirely their own.
They are so driven to make it better. So excited to share it with others, get feedback. They love seeing their friends use it. And so, that's created a ton of the drive and it's also enabled children to sort of build their own, go out and look at social problems, stuff like that. And so, we've had AI club teams that have won like 18 competitions in the last year and a half, all custom projects, all different, all built on things that they've learned.
So that's really what the technology powers. So what we've kind of brought to bear is sort of our understanding of AI, how to extract the essence of it, how to connect it software wise and technology wise, that's what our engineering team works on. Is how to make that connection and how to create the automation and then slowly expose it.
So long-winded answer but hopefully that answered your question. - Nah, absolutely. An interesting point you made there about the layers, that's a great way to keep them interested and not feel more importantly, maybe disenfranchised with the process of a sense that they go, oh, this is for a kid, but this is not for my level. And so, it kinda naturally adapts if you will, I guess, right? To their knowledge level.
So I think that's pretty creative because yeah, sure, no one wants to say, oh, I don't wanna use that tool anymore because that's for kids when it's adapting to them. So that's something but you also had a good point about children and how the parent was frightened of that conversation. Everyone gets worried about Skynet, perhaps that's not something we need to worry about, it's the children with these conversations because they're outpacing us on all this stuff.
I know my dad feels the same way. So the question again, this is going to leak from that, it's to both of you again, and I'll start with you Elnaz. What are some of the challenges? And you actually touched on this a little bit in your previous answer but we wanted to elaborate on it because it seems like a space that's pretty important and challenging for both youth. You know, what are some of the challenges working in the EdTech space especially where the youth are involved? If you could elaborate on some of the things you touched on earlier about that? - Sure, definitely.
I think one of the major problems in the EdTech space is, it's very slow. And maybe that's one of the reasons that we decided to deploy our technology and start it from home and then gradually from home, get the robot into school system. But the good point is, during pandemic and after pandemic, we see actually a lot more teachers are more excited about innovation technologies, getting new products into school system. So I feel like it's changing gradually but it is traditionally slow.
It is one size fits all and those are the areas that they're challenging but when there are challenges, that means you can come up with innovative ideas, bring incredible inventions into school system to really change it. So that's why I always think it's an opportunity for everybody and all of us in the education space. But also, when we work with children, it's a matter of privacy. It's a matter of data and so many other limitations. I also remember when first we launched ROYBI, it was exactly at the time that Netflix had "Black Mirror". So we had to deal with so many people's challenging questions saying, "Oh, so your robot with AI "is going to take over my child's life."
And you had to constantly educate them that no, it's not gonna be like that. The robot you see in the movie, it's totally different. AI is not even in that space anyways, you know? It's less intelligent than that.
And I think the other challenging part is, we have to continue educating people to understand AI is really here to help their children to focus on their individual abilities, interests. It can really help people to do more intelligent assessment, create a lot of programs that Nisha was mentioning. So there are a lot of opportunities but I think it takes time. But then, in order to make past these challenges and also make education accessible, I think we need some technologies like Edge which you know, it removes the connection to internet, so this robot can go anywhere in the world. They do not have resources.
They do not have internet access and it can just teach the kids as a tutor. But of course, not replacing teachers because it becomes like an aid, a support to teacher. They can put their programs in a product like ROYBI and be able to access those areas and children that they really do not have access to these type of technologies. - Thank you. - Yeah. - Nisha? - Oh, yeah, so I'm thinking of, yeah, so the challenges of kind of AI in EdTech I mean, I would sort of echo some of the things that Elnaz said.
I think EdTech has actually in overall looks like it has benefited. It's hard to say that it's benefited but there's been a resurgence of EdTech because of the pandemic. And also I think generally seeing more of a kind of a growth in EdTech because fundamentally, I think education systems are changing. Now, pandemic forced a level of change that wasn't natural.
It was happening anyway as people rethink what it means to learn. You know, it was happening very heavily in K-12 in Asia and more so in the higher education systems in the United States but the pandemic just threw all of that into overdrive. That it forced everybody to think about how else can I do this? And then the real question will become now, will we go back? Can we go back? Because frankly, when right now it's not entirely clear, we can't go back even you and if we wanted to. So EdTech is definitely sort of growing in that sense. The challenges with kids, privacy is definitely a sort of a big one. One of the things that we have done which is, I mean, the privacy thing is just a given, it's something you have to deal with very carefully and important.
Additional things when you're teaching AI to kids, it's a little known thing that you might not realize is that a lot of public data sets are not suitable for children. So it's interesting because AI lives on data and without data, you can't really do AI. Now, if you go and look at data, there's a ton of data sets that for a variety of reasons, particularly the ones involving language for example, are just not suitable for children.
You should not be putting that kind of language in front of a child of a certain age. And so, one of the things that we have done is actually we've aged curated. We have like 400 curated datasets which are kids safe by definition and things like that. And these are some of the things that you have to do when you actually start doing AI literacy for K-12 is you have to look at stuff like this. In addition to protecting their privacy, it's also what are they allowed to see and what do they get to work with that is kinda safe and protected for them? So these are some of the things so that, we've actually built a bot, our bot is called Chai the AI Bot. It chats with kids and answers questions about AI and helps them with their homework and stuff like that and it's got a safe filter built in.
And we have had many kids try to get it to say things it shouldn't, you know. Because they'll get a kick out of it and then they get through all. It basically says, like, my teacher tells me I should not be answering your question, you know? And stuff like that, to just get them off of the thing of trying to get it to say things that it shouldn't and stuff like that. So there's a lot of these kinds of things when you do teaching with kids and so forth. And so, those are sort of some of the things that come to mind about focusing on AI in EdTech.
Now, talking about AI awareness and AI literacy in general, you know, I've definitely experienced the same thing that Elnaz has said which is, people tend to see AI as something that is either all knowing or something they should be afraid of. And the problem is that in commercial products, AI has intersected with privacy in a way that is actually slightly dangerous, if not dangerous. But it's not the AI that's making it dangerous. So if you decide to tell a digital assistant all about your personal life, it's not the AI that's gonna get you into trouble with that. It's the fact that that company is going to record that information, right? And so forth and then do something with it that you probably don't want.
And so, part of it is we do a lot of education sometimes because we feel it's important, sometimes because it's necessary, but we try to help people understand that the technology underlying is not at fault but you do have to be extremely careful about how you use devices that have this technology. In like one exercise that I sometimes tell children of middle school is, come back tomorrow and tell me if your digital assistant knows your name and whatever you do do not tell it your name. I wanna know if it knows, right? And then they come back with, okay, Prime seems to know my dad's name.
It seems to think we have an account, right? Doesn't seem to know my name which means it hasn't figured that out yet and we have those kinds of exercises. And so, yeah, education is a very big part. I personally believe that AI awareness is critical. That AI literacy is critical but teaching people how to be literate means they should know what it is. They shouldn't be afraid of it for what it isn't. But at the same time, we don't want to think it's benign because it's not.
So, hopefully that answers your question. - No, absolutely and more. I mean, I think that's a really good point about the responsibility aspect that's on the user maybe. And these are not just awareness of what AI is and how prevalent is but the responsibility that comes along with using it and then not sort of putting weight of that responsibility on the AI itself but the user. Yeah, I thought you made also a really great point about the data set.
That's something I had thought about which is, you curate the tools or as you literally peel back the layers based on the seniority of the student, let's say, but the data set itself is curated based on the fact that they're students. And that's a really interesting point because I think a lot of people also think about AI but do they think about what constitutes that AI is the data, the input, right? That input is paramount for AI to exist and be able to do what it does. So that's a really interesting point.
And I thought it was also neat how you guys both had more parallels. So I love talking to you both at the same time because you both talked about different things that are exactly the same thing which is, you know, talking about areas in the world where you may not have internet, right? And that was Elnaz, Nisha and then you mentioned that idea of like bringing technology to those folks that may not have it. And so, that idea that, oh, COVID, excuse me, COVID, my brain had a moment. The COVID changing sort of the paradigm of how we're teaching, right? So again, COVID all across the world but that is, yeah, how do you bring these tools where you may not have internet? Or maybe the students now aren't going into a classroom. You have to bring it to them and adapt to where they are. So I thought that was really neat 'cause you both touched on it but in totally different ways.
So again, a lot seems similar so I'm gonna use and talk about but so a question again for you both and you already again, touched on it, beat me to it, I love how it sort of interspersed into your answers already but you know, I think we'd all agree that AI is considered as I mentioned in the very beginning emerging technology but that being said, what and again, you guys have touched on this but I'd love to hear more as what I'm sure audience, what emerging technologies within AI do you guys see on the horizon? I think Elnaz mentioned Edge but I think our audience might wanna know what you all think is the next thing within AI. - That's an interesting question because when you think about the future, 20 to 50 years later, it's going to be totally different than what we see today. I remember even sometimes you have conversations with my mom. She was born and grew up in Iran and it was a village and she was telling me for some time when she was a child, they didn't even have a TV. So from that time and she's not old you know, from that time to now, with this much advancement in technology, I can't even imagine what's gonna happen 10, 20, 50 years later, right? But some of the industries that I see they're going to really utilize AI and bring some incredible innovation aside from from education which I believe is going to be the next big thing.
I say, probably like entertainment. You know, now we see the use of AI and machine learning in entertainment industry. Whenever I go to my YouTube, there's so many recommendations but also thinking about it you know, now you can actually build characters with AI. What if we can create films with our own customized characters and actors, actresses and create our own films and within the entertainment industry just using AI? It is totally possible. And of course, adding virtual reality to that, yesterday there was an announcement from Facebook that now they're using VR for remote learning and I look at it, I'm like, "Oh my God, this is totally the future."
And then of course, some other areas for example like medicine, I've been hearing a lot of advancements using AI in medicine. There are so many companies that use AI to detect some diseases, cancer, heart disease, especially much, much earlier ahead of time. They are building some treatments, earlier treatments just because AI can recognize a lot of things ahead of time. I'd say transportation, we got autonomous driving and I'd probably think next time 10 years later or five years later, we request a car from Uber, it's just not gonna have a driver. (chuckles)
So it's just really amazing and exciting to think about the future. And even our daily tasks you know, reminders, shopping and everything could easily be done by AI. And I just can't wait to see what types of ideas and innovations people are gonna have.
- Absolutely. So I thought maybe like one way to kind of illustrate what the future of AI might be. Let me tell you some of the things that middle school kids are building now. You know, I've had middle school kids detect breast cancer, Alzheimer's, glioblastoma, various kinds of diabetic retinopathy diseases. They've simulated self-driving cars. They've looked at climate change.
They've recognized pictures of currency to help the blind. They've created a chat bots. All of these things are now within the reach of the average middle schooler. So you can imagine what is within the reach of an adult, right? If that is what is in reach of a middle school. So I think that probably the top level things that I would say in terms of where AI seems to be going as well as what it means is, I think we're increasingly heading into a future of one personalization. You know, personalization as in everything that we interact with from the drugs that we receive to the prices we get quoted for stuff is pretty much now just for us.
There's no standard price anymore. Like when you go to call an Uber, that price is increasingly for you. And the more they know about you, the more that price is about you. Right now it might just be time of day and where you happen to be standing but later they will figure out more about you and then that price will be for you and the guy standing next to you is gonna get a different price because they know something about you and what you're willing to pay. Now, depending on how you look at it as both good and bad, you know, for people like me who don't like to negotiate, I don't like this at all.
I don't like the idea that I don't get a standard price but if you're the kind of person who negotiates, it might be good that you got a different price and particularly if you've got a better one. So that personalization and this is creating massively value in things like healthcare. You know, personalized medicine, particularly when it comes to things like cancer and so forth you know, half the battle of cancer is whether you respond to the medication. And if you don't respond to the medication, there pretty much isn't anything that can be done. So if anything can find the medication of all the options that you're likely to respond to, that has a material impact on your life.
So this personalization trend is going from things like shopping to medicine, to just about every aspect of anything. There's personalized learning plans being created by tech companies to help you assist K-12 students with their learning. Everything is just becoming custom because the algorithms can now handle it.
So that's sort of one important area. I think the other area which is sort of related but is interesting is also kind of tracking. So in different countries tracking is becoming a very interesting and challenging problem.
So like one of the things that we teach our older kids is we ask them to find out how many places in the United States have banned facial recognition used by law enforcement, stuff like that. And those things are being used in many countries outside of the U.S. in ways that we might not consider to be acceptable. And so, all of these sort of the swirl of information, personalization in some ways the positive use of it, tracking is kinda like the negative use of it, if you will.
So that's one thing. The second I think major trend is sort of the up-leveling of the human decision. So I don't know if you guys know this but AI can now write code and has been able to write code for quite some time. And so, now what does this mean for the engineer? Doesn't mean you replace the engineer, it just means that the engineer doesn't have to worry about the nitty-gritty of the code. They should now spend a little more time thinking about what their user wants from their app.
'Cause before that they were worried about little things like why isn't this line of code working. You don't have to worry about that anymore. AI created a line of code that works and stuff like that. And so, this kinda brings us back to the point that Peter made about responsibility and I think this is really important is, you know, one of the things that we try to teach kids as part of AI literacy is helping them understand their responsibility. Like, AIs learn what they're taught.
So whose responsibility is it to make sure that they learn the right things? You. Similarly in a world where AI does a lot of stuff, the human has not abandoned responsibility. The human's responsibility has simply changed. So I think those are sort of the fundamental trends because AI is showing a capacity to learn from generalized information like the internet.
It's showing a capacity to answer general questions, write code, build things by itself. So it's not just that it's solving problems, it's building things that will solve problems which is another level. And I think that the line between that and Skynet is responsibility and guidance not mechanics. So that's where I think AI is going.
- No, that's a good point about code, a lot of people knows, I could have used that in college. I could sit until 2:00 a.m. in the lab trying to run an email client job, I very much would appreciated any help getting that done but that's a really good point. And to the point you made about individual, customization, I mean, they're even using it, right? I saw one article with automotive manufacturer company was using it to have AI on its own to figure out a very individualized problem for the wheel housing or gear system and it sort of crafted a very alien, they called it, looking way of connecting the metal from the axle to the wheel and things like that. It was very individualized to that car, to that design but they just unleashed AI on it and it sort of figured out a very custom one, not for any other car, not for any other design but just for that design. What would be the optimal with the stress amplifications and the rigidity and all of this other stuff.
It's very interesting when you see targeting things to very specific cases and I think you're absolutely right about that. I'd be remiss if I didn't ask more of a patent question, I have to get this in here and I will tell you the AI Working Group that I'm on help me with this one. So as you may recall, we did talk about this briefly in our previous chats. The USPTO have recently ruled the AI systems cannot be credited as an inventor in a patent. That under current law, only natural persons may be named as an inventor in a patent application.
What do you both think of the recent news globally where in certain places it's been determined that AI can in fact be credited as an inventor? You know, what role do you see patents playing in the field of AI? I'm sure the audience would love to hear sort of your thoughts on that. - It's quite interesting actually, when I heard the news I talked about it a lot because based on the current data, client information and progress of AI, I have to say it's relatively low intelligence right now. And it still needs a lot of human interactions.
So when it comes to filing your patents, you know, based on how I see it is, the inventor is pretty much same as the software or the AI behind it. It's like the same thing. So naming for example, AI as an inventor or being somehow, of course it can be part of the software, right? But I just think it's quite an interesting subject. But I also feel it depends on every region or country's culture because for example, in a lot of Asian countries like in Japan, a lot of people they think objects actually have souls. So they are a lot more receptive towards objects and seeing them as part of their lives.
So if for example, in Japan says AI is an inventor, separate inventor, I wouldn't be surprised but from my point of view, I still think it's the human, the engineers or the people behind it that can actually originate the idea. And of course, build the foundation for the AI programs to build things upon it. That's my opinion now, I don't know if it changes five, 10 years from today but I don't have any facts to say, AI would be a different inventor on an application.
- Sorry about that. So, yeah, so I think, I guess my take on it is, I mean, there's always this fascinating sets of news about AI that makes you think and this one definitely makes you think and in some ways it's good that the question is being asked. I feel like naming an AI as a patent inventor does not serve any useful human purpose. The purpose of naming humans as inventors is partly for their credit and their careers and making sure that the company being an assignee is for their own patent protection and things like that.
It's not clear who benefits from, because the AI doesn't care, at least yet, you know. The other thing that I think raises a really interesting question is that AI can be replicated instantly. So the brain of an AI is literally a file. You copy that file with any tool you have on your computer, you got another identical AI and I'm not sure how you could argue that the first one got the patent and the second one didn't when they are exactly the same but it's not like we can replicate human brains and have this argument. So the whole thing just feels like it's just a little weird and somehow just doesn't sort of stand up to the day-to-day workings of how these things are used.
You know that's said, I think it is interesting that different countries are approaching it differently. That's probably the bigger takeaway here is that every country is approaching AI differently in some combination of their perspective, their culture, their laws are sort of intersecting with this. I do think there are also other interesting patent issues around AI.
One of the things that I can tell you that companies like myself, even not just my current company, but even my previous one kinda struggled with is, AI is very much tied to open source. And open source is very much tied to certain kinds of legal licenses and that's part of the reason AI has grown so much is because of how open it is. And then the real question becomes, what is it that you're patenting? You cannot patent an algorithm because it probably didn't belong to you, it was probably invented in some form decades before you were born. However, that one model that you spent years training, that took in all your data and all your things and was tested and tuned, is a real source of IP and as far as I know, I'm not entirely sure that set of numbers can be patented. So I think there are lots of questions like that as to how, right now I believe most companies are protecting that simply by not letting it out the door.
Like we don't know the numbers that make up Google search algorithm and we will never know, right? And that's how it's protected but that is the thing that makes a difference between the core question and the quality of the answer, it's not the algorithm. And so, how is that protected? It's a hard question. I think it's an open question for the future.
- I think that's a great point actually. I had not thought about that point about easy replication. You're right, it's just basically, you take this file and you clone that. Very interesting, discussion for a later topic. So, we're just about out of time Elnaz and Nisha, I just cannot thank you both enough for sharing your stories and your palpable passion for AI especially AI literacy with us. Are there any concluding comments that you each would like to share with our audience? - Sure, first of all, I want to thank you and USPTO for inviting us here, it was quite an amazing conversation.
And I really want to encourage the audience to really start thinking about AI and really working on creating some technologies and bringing inventions and innovation to this space. There's a huge, huge opportunity. And I also want to invite the viewers if they are interested in some collaboration with ROYBI, please do reach out to me, the best way would be LinkedIn. - [Peter] Thank you. - [Nisha] So similarly, I think thank you so much for having us. Thank you, USPTO.
Thank you, Peter and Dennis, I think this was a wonderful discussion. You know, basically, I would encourage everyone to really think about the AIs that are around them. Think about what they understand, what they would like to know, how can they become more aware.
If you're a parent, what do your children know? How can they become more AI literate? You know, how can you kind of ensure that AI literacy is a part of their lives? And if anyone would like to collaborate, we collaborate with lots of organizations both non-profit and for-profit organizations to bring AI literacy worldwide. So, LinkedIn is a great place to reach me as well. Just drop me a note if you'd like to chat about AI literacy. - [Peter] Thank you both.
Once again I like to thank Elnaz and Nisha for spending their time with us. I'd also like to thank everyone in our virtual audience for joining us on this panel. And I'd be remiss if I didn't also thank USPTO's Natanya Ferguson, Dennis Forbes, Sean Wilkerson, and our amazing conference services team for making this event possible. If you're interested in learning more about what USPTO is working on on the AI front, please visit the AI page that we have which can be accessible via a link that's right on our main website. So just go to uspto.gov, you'll find the link on there
for our AI landing page with a bunch of interesting information. So before I head off, I just wanna say stay safe, keep innovating and thank you both for everything.
2021-11-05 23:43
