[MUSIC PLAYING]

VOICEOVER: Is this the accessibility moment we've all been waiting for? Speaker, Saqib Shaikh, Project Lead for Seeing Eye, Microsoft. Moderator, Mike Shebanek. Product and Market Strategist, Be My Eyes.

MIKE SHEBANEK: Thank you, Karae and Ross. Really appreciate it. We are so happy to be here at SightTech Global, and we want to especially welcome everyone and those who may be joining for the first time. So, Saqib, great to be with you. I can't wait for this conversation. I've been looking forward to this for a long time. And I'm going to start with a question I've been wanting to ask you for a while. I don't think I've ever had the moment to ask, which is, you've been at this AI thing for a really long time, like since 2017, I think it was, when you created and kind of launched the Seeing AI product at Microsoft. And so I'm just personally curious, I'm sure a lot of other people are too, what was it that caught your attention about AI to begin with and how it could be used for people with vision loss so much earlier than everybody else around? Like, you were one of the first people doing this work, and I'd love to know kind of how you got into it.

SAQIB SHAIKH: Well, thank you for the kind words, Mike. I guess, you know, my journey with AI goes all the way back to college, where I did a degree in artificial intelligence before it was cool. And it was just, you know, something that really fascinated me, this idea that computers, yeah, they could be programmed, but they could learn themselves. And then it was actually 10 years ago this year that I started working on Seeing AI as a hobby project. And then within Microsoft, it just got traction. There was like a lot of great colleagues helping out, and eventually we got the CEO's attention. And it was all driven by this idea that as someone who's blind, what if I had an AI being the equivalent of a sighted guide, even when someone wasn't around? So, you know, I sometimes talk about a friend sitting on my shoulder, whispering in my ear, looking around, and it's a funny little mental image. But that's what's taken me all these years. And like you said, 2017 was when we first launched Seeing AI to the world, and it's been a great ride ever since.

MIKE SHEBANEK: Wow. Yeah. And so early on. I mean, even as you said, even before Seeing AI, you were already into the AI and kind of what it could do for people. And then now, of course, it's exploded. AI is talking; everybody's talking about AI. Of course we are too. It's everywhere. What does it feel like to have the world sort of catch up to you? Like you knew about this way before everybody else. And people were probably like, what, what are you talking about? How really is that going to work? And of course, Seeing AI is like one of the most popular apps in the blind community. So what does that feel like to have everybody now part of the conversation?

SAQIB SHAIKH: Oh, it's great because what it now means is that, you know, it is a part of the general, you know, the general consciousness. And so it means technology is getting even better. And laying out the possibilities of what's going to come next. So we see this going back decades and decades that technologies that were first created for people with disabilities later become part of the mainstream. And there's dozens of examples from talking computers and speech recognition. I could lift so many. Yeah. But again, this idea of using cameras to see for someone who's blind and so forth, it's been there for years. And so it's great to see. And it's great to be able to feel the investment now and to think what's next.

MIKE SHEBANEK: Yeah, for sure. So that leads me to a question and really kind of the title of our session. So is AI just sort of another step along the way of a whole laundry list of kind of improvements to accessibility over the last? I mean, you've been working on this for decades, so have I. Or as the title of our session suggests, is this the accessibility moment we've all been waiting for where things are suddenly great?

SAQIB SHAIKH: As always. A bit of both. I think right now what we've had is this huge leap forward, but I also don't think it's problem solved. There's still so many challenges to be solved. And hey, that's exciting. That's why we're in this space. But absolutely what they call generative AI, which is what's really been catching people's imagination recently, it's a huge, huge leap forward. And it is opening up so many opportunities, and we haven't, well, yeah, we just scratched the surface of that so far.

MIKE SHEBANEK: Yeah. Okay. So you mentioned something called generative AI, and I recognize there are probably people that are in our audience that are totally up on all of the AI lingo and language and all of the nuances of it. And you're kind of curious about where that's taking us, and for others that might be their first introduction to what AI is and how it works. So maybe we could just for a moment, just sort of stop and talk about like for the average person, what is AI and what is generative versus sort of what we think of as just AI? Like, how are those two things different?

SAQIB SHAIKH: Yeah. I guess there's been, you know, many iterations of AI over the years, but this recent leap forward, which you might have seen in apps like Microsoft Copilot and ChatGPT, is this idea that the AI can make up answers. It can generate answers to questions it's not ever seen before. So it can create a document for you. It can consume lots of information and make brand new content or reformat or restate things in a totally different way. And that leads to really huge opportunities in terms of personalization. And if you think about accessibility, that is all about personalization, that each and every one of us is different. And if we want to include everyone, then AI that can take information and personalize it to your individual needs is. Is really transformative.

MIKE SHEBANEK: Yeah. I think that's maybe one of the fundamental differences, right? So assistive technology in the past was kind of, we, it got created and everybody kind of used it the same way. It kind of answered the same way, worked the same way. And AI sort of promises to be very personalized and unique and adaptive to the person using it.

SAQIB SHAIKH: Yeah. And I, I, I'm quite excited to see where this takes us in terms of systems that are not one size fits all, but unique to each of us.

MIKE SHEBANEK: Yeah, for sure. So, so, so, uh, obviously AI to learn all these things and be adapted, uh, it has to get some data. It has to kind of be smart, right? Where does, how does it get smart? So where's all that data coming from? Uh, I, I, obviously there's some things on the web and the internet, but, um, like it, it, what are some of the other ways that it's getting this information?

SAQIB SHAIKH: Yes. So going on to your previous question as well, the work we've done on seeing AI has been sort of, you know, describing images or asking questions about. And it can right now say, even write a whole paragraph for you about what it's, what the camera is seeing. And you can ask questions to find out any information you want so that it's just that piece of information. But as you say, where's this information coming from? Are those descriptions accurate? Are they representative of, of you as an individual? And so right now, a lot of that information. Is just coming. It's the same information for everyone where we have these large, what we call models, large AIs, which are trained the same way for everyone. But some of the work we've been doing is looking at, okay, through a partnership with Be My Eyes and others, can we get data, which is specific to people who are blind and therefore create systems, which understand the lived experience of the blind community even better and can serve this community even better. So. That's definitely an area of focus because there are things which are specific to this community, which might not be captured in that more general set of data.

MIKE SHEBANEK: Yeah, and I, and I think the, in the past, you know, the data has come from like, you know, social media or websites or PDFs online or scan documents, things like that. I think one of the interesting places that is coming from now are sensors, you know, we'll talk about this more, maybe a little bit later on. But, um, the fact that you can carry. Around a wristband that captures your heart rate or knows how far you're walking or GPS is of where you're standing or where you're going. Um, and all of a sudden now really make use of it and kind of completely new in different ways. And those, those sensors, if you're wearing them are obviously very personal to you and what you're doing in the moment. And, uh, yeah. So what are your thoughts on kind of where that's going or is that, is that where most of the data will come from in the future? Or do you think it's going to be more of just, you know, putting books online and, and, and, you know, people talking back and forth and learning how they, they communicate.

SAQIB SHAIKH: Knowing the date is going to come from everything. Like you say, right now, we talk a lot about text-based data, but with what we call multimodal models, it's like different types of inputs. So it could be video, audio and so forth. But as we get into real-world systems, um, yeah, it can be all of your senses in terms of, yeah, what a machine can see or the movement, or even like it. A, for healthcare, um, your heart rate and so forth is really important for those applications.

MIKE SHEBANEK: Jason D's response was actually quite clear at UFSTEL patient Yeah. Uh, I think the thing that's really amazing and creating some of the urgency around this is how fast things are moving. Like you've been in technology a long time. I've been in technology a long time, and we're kind of jaded and used to things like, okay, next year it'll be Covid-5D. and stronger; and year after that, I'll probably be a little faster than that. But honestly, and especially for you, since you've been focused on AI even longer than I have, the speed at which AI is transforming things and getting smarter is staggering to even the most jaded people in technology – it's incredible! And for me, like the breakthrough has been, and these are technical terms that I'll share with people who are in the audience. You know, there was something called sort of supervised learning where you kind of had to manually look through the data and teach the AI what it was and how it was categorized and what it meant, and all these sort of things. And that became sort of semi-supervised, and now we have this thing called unsupervised training – I don't know if you want to share a little bit more about like how that works and why that's become so explosive. But to me, it's incredible because machines don't turn off; they'll just start learning on their own 24/ 7 and and never stop, so the Speed at which they're learning is just so incredible and it's it's it's it's it's it's it's it's it's it's it's it's they're getting smarter and acclimatizing this data is absolutely incredible. But that obviously has impact on sort of where we think the future is going to go and why it's urgent. Do you think that was kind of the magic moment where things changed from 'it's a new technology that someday will get better' to 'this is like in a good way, like a runaway train, just getting better, faster, like in ways we can't imagine?' Because, like, what was your thought when we went from like ChatGPT-2 to ChatGPT-3 and then suddenly four, and everybody's heads exploding at how good this thing is getting and how quickly?

SAQIB SHAIKH: It's a really exciting time to be in this space because you're really transforming, yes, the work we're doing in seeing AI for people who are blind, but then, you know, across all of industry and society. And every few months, there's a new breakthrough, which makes something, something new possible. And it really makes you reconsider what's possible. It's this moment where it's like, OK, it's not possible today, but maybe it will be in a few months or next year. So that's just really, really exciting. An example of that is I: if you'd asked me a couple of years ago, I would have said describing a video for someone who's blind was years away. But in seeing AI, we just launched a feature that lets you describe videos. And in that way, it's just so exciting, because what I do in terms of my work is really thinking about, OK, as we learn about the needs of the blind community and what the latest science makes possible, how do you bridge that gap? And so many new things are being made possible every few months. And yeah, that is driven by many things like the new algorithms and technologies, more powerful, supercomputers in the cloud. And now that we have the whole industry focused on AI, it really means that, you know, those investments will help this community as well.

MIKE SHEBANEK: Yeah, I think that's the key. It's not just an accessibility thing where it's struggling for resourcing and people to focus on it. And but we're leveraging this giant, massive industry-wide investment in AI, and we're able to to integrate now assistive concepts and technologies that will can apply and really help people in ways that didn't exist before. You mentioned earlier something called multimodal. And for people who don't know what that means, if well, effectively, it just means that it can take on different types of data. So it's not just text, right?

SAQIB SHAIKH: Right, multiple modes of input. So whether that's audio, video, text, or others.

MIKE SHEBANEK: Yeah. And so you talk about video being added to seeing AI. And I think the so being able to do photos and ask what's in a photo being able to interpret what's in a video. Obviously, text, you can get to human voice, can you recognize what I'm saying? Obviously, it allows us to have the interaction with these things in new and different ways, besides a traditional keyboard, right? And that's a game changer, at least from my perspective. I wonder if you have the same thought, because I know for a lot of people that I've encountered that use a screen reader, learning all of the amazing number of keyboard shortcuts to be able to be not just, you know, functional and use it, but use it fast and efficiently and be really good at it is daunting. It's daunting for a lot of people. But the fact that you can now use AI and effectively just talk to it is kind of opens this up in a way that I think maybe I can use this analogy. You know, in the past, we've had like laptop computers with screen readers that have been amazing and still continue to be amazing for many people. And then we went to like smartphones where suddenly all you needed to know was tap, double tap, right, and drag. And all of a sudden, so many more people could get access to not only sort of assistive technology, but general technology, like they had a smartphone in their hand. They could, you know, order food or get an Uber or find out what the weather is going to be or, you know, do my clothes match, whatever those things happen to be. Just AI feels like that to me. And I'm wondering if it feels like that to you. And that's why it's becoming such an urgent thing to talk about, because suddenly the net gets wider. And so many more people who thought either I can't do this or it's too hard to do this or it's too much to learn can now actually take advantage of not just AI, but, you know, the AI can help them take advantage of all the other products and services that might come along.

SAQIB SHAIKH: Well, yeah, AI is an incredibly powerful technology, but it is a technology, it's a tool. And so, you know, every day I remind myself what really matters is the experiences we're creating. So I see my work at the intersection of creating novel experiences that really help people in their daily lives, powered by this really powerful AI. So if you think about this, then on the one hand, maybe you will speak to your computer or your phone or your hands-free device, but that's not always the thing that people want. And it can actually be quite inefficient or socially not acceptable. So we talked about personalization. So I think these experiences need to be able to cope with all types of users. Someone who wants to be productive at work, but then also someone who might be intimidated by computers and smartphones. And can we bring more people into the fold of, you know, using these powerful tools? So you want to include everyone and the way different people will interact and even the same person at different times in their life will be different. Will it be gestures? Will it be tapping physical buttons or something else? And this is where this idea we mentioned earlier of one size fits all, I'd love to see that totally go away. What if the AI knows about you, your environment, your current needs, whether you're at a computer or interacting with the physical world, and then gives you the assistance you need at that moment in time. I think this is a really powerful concept and it's something where, you know, we're still quite away from, but working towards.

MIKE SHEBANEK: I think this is a really powerful concept and it's something where, you know, we're still quite away from, but working towards. Yeah, it's amazing contextualizing of it, like you said, like you have it when you need it, it knows about you, it knows what you want and how you want the information presented, but this idea, I mean, and this is kind of fundamental of like seeing AI and of course I'm working with Be My Eyes and we're working on things like the Ray-Ban metaglasses where you're simply wearing glasses and your hands are free to carry things or pick up things or interact with the world while you're utilizing this service to find out what's in front of you and ask, 'What do you see in the camera?' You know? Can I get information about something that I just happened across in the moment? It's sort of like having the phone, but you have to take the phone out of your pocket if you're using a cane and holding a bag that you can't get to your phone. So this idea of being able to have things hands-free and in the moment and sort of instantly when you need it is really compelling and something I think is not in the distant future. It's literally now, like that's kind of amazing.

SAQIB SHAIKH: Yeah, it is now. So, you know, we've launched some partnerships with some companies making wearable cameras. So you have the hands-free experience and we're definitely working with others, but it goes to this bigger point of you want the AI, you want the technology to vanish. You don't want people to need to think about technology. If the technology is good enough, you don't even know it's there. And that applies, you know, to being in the office, working on a laptop, but it also when you're walking down the street. So can we have something maybe that you wear? Or maybe it's part of the room. Maybe it's just something that's on your PC by default. And then the AI can make use of all that information, all the sensors and convey that to you in whatever way is best at the moment in time. And as we said, personalized to what you need. But yeah, who knows what shape that will take? And ideally the shape of it won't matter because it will just be so natural. But again, that's a way away. But working towards it.

MIKE SHEBANEK: Well, we've had our moments, right, as an industry, and I think immediately when you said that, I thought, oh, right, we had this thing called the Humane Pin, which was like a lapel pin or something that had sort of an AI built into it and some sensors and things. And it didn't have a good landing, honestly. Like, I think it's well known now that it wasn't well received as maybe as it ought to have been. But I'd love to know your thoughts on like wearables. Do you think it'll just be that we just have AI on wearables and we don’t need laptops? And phones and other gadgets are like, would you do that or would you say, no, no, no, I want to have my laptop AI and I want my phone AI. And then if I have a wearable, that's cool, too. Like, how will you use it?

SAQIB SHAIKH: And I really see that, you know, theoretically, each of us will have maybe one AI, but maybe also each of us will have many AIs for different tasks because, you know, if we want advice or guidance, we'll go to different assistants in the real world. But yeah. The AI pin you mentioned is a great form factor for someone who's blind because it's just so small and unobtrusive. That product did not, you know, apart from that product, I like this idea that sometimes you will wear a camera or a microphone or another sensor, but there might be other times when you're at work and it's just part of the meeting room. And of course, this would be conserving privacy, security, all these great things. But then, you know, why, if the AI is able to know, if you want to know who's in the room, who's around me, where am I in the building? It could be that the thing you wear is the sensor, but it could also be something that's seeing you and guiding you from a third-person vantage point. So, yeah, lots of exciting possibilities and continuing to explore what we can do. And as you said, because the industry is now moving so fast and there are all these mainstream applications coming on board, it will open up new opportunities because I do feel that the early adopters in this blind community are often years ahead. And so we're trying to work with the scientists to see, okay, here are the needs of this community. And even if it's some way away, what can we start putting in place now?

MIKE SHEBANEK: Yeah, it's funny as you were talking, I was thinking. Wouldn't it be weird to have like an AI in a wristband and an AI in glasses and an AI on your smartphone? And you could ask one AI to ask the other AI what they thought about the third AI. You could have this great, wonderful AI on as you're walking along, having this wonderful AI conversation. You could listen in and go, no, I don't think that's right. What do you think, Mr. AI one? Or how about you, Mr. AI two? It'll be really, and I think that just highlights like the interactions we're going to have and the thing, and especially like little kids. Imagine if you're like a five or eight year old kid in this world where they have all these AIs and they're going to think of stuff like this. They're going to do crazy things we never imagined and ask the AIs to help you. Well, why don't you help AI one learn about AI two and how I like this to be done or how I like that answer. It's going to be a wild world, isn't it? It's going to be really, really different.

SAQIB SHAIKH: It is, and I think all the AIs will just talk to each other behind the scenes. We won't even know it's happening. But, you know, yeah, the AI that knows about what's right in front of you will talk to the AI that knows. About I know you're all the things you've historically done, you know, whatever it is, and yeah, you'll just get assistance. And again, I think there's a lot of excitement in the general population about these things. But bringing it back to sort of the assistive application, what we're really thinking about is, okay, there's a task you want to complete in your life right now. Like, you know, something you want to go and find something you want to go and read something you want to whatever it is. And. Yeah. Or an interaction you want to have. So you really want all the tools with all the greatest technology, whether it's hardware, software or, you know, even robots one day. Right. All coming together to assist you in whatever way is best for this individual in this moment in this space.

MIKE SHEBANEK: Yeah. Well, you brought up robots, so now I can't help myself but ask you about robots. There was a conference, I think it's just in the last month, called the 20**24 World Robot Conference. Where they talk about the very latest in many different types of robotics, but especially humanoid-style robots. Tesla talked, I think, a while ago about their Optimus humanoid robot. Amica, I think it's called, and Ozzy from Engineered Arts has been around for a while, really well known and famous. And I think what caught my attention, first of all, aside from the capabilities, but I thought they're like, no, these are exotic, right? They're prototypes. They're going to, you know, some color. A company might be able to afford one, or maybe they'll cost $100,000 US or something. Like, these are effectively around $14,000. This is less than a car. There's bicycles that cost more than that if you're into racing bicycles. So, these are real, these are now, and these are relatively affordable. Now, we're really early on. So, at some point, that price is going to become, you know, I don't know, $1,000, who knows what it'll be, $1,000 or something. But the functionality they bring in, the fact that you can enable it with AI. And this is another way to interact with AI, besides wearables or smartphones or computers, suddenly gives them capability, right? Because now it's not just, hey, can you give me an answer? Can you tell me what's around me? It's, hey, could you go make dinner? Hey, could you get the door for me? Hey, could you walk me down the street to some place? I've never, if you wanted it, you don't have to be, you could obviously be an independent if you wanted to be. But, like, robotics and AI in combination is pretty amazing. And I don't know, especially in the blindness community, if we've talked much about that. And, like, what are your thoughts on, would you like to have a robot that's AI-enabled that you could have as a, when you wanted it, either around the house or outdoors or wherever?

SAQIB SHAIKH: I would love to have loads of robots, because.

MIKE SHEBANEK: It's your robot army, yes.

SAQIB SHAIKH: You know, I don't think they're going to be humanoid robots. They're going to be gadgets. Our gadgets are just going to get more and more advanced. And I think we're going to have robots. We really need this science fiction idea of humanoid robots to push the industry forward. But by the time it comes to something that you buy, it's going to just be a device that does something special for you. And to your point, I do think the medical field, and then that means people helping people with disabilities, will be one of the early adopters. So maybe it will be a device that helps me avoid obstacles. Maybe it will be a device that goes and helps me pick things up. And we're already seeing this idea that AI can not only tell you about stuff, but act on your digital world. Like when you're at your computer, it'll, you know, create a document for you. But then you could take that to the physical world, where in the physical world right now, you know, with the cameras and microphones, the AI can sense and tell you stuff. But it can actually do stuff for you. So instead of guiding your hand to pick something up. Yeah. You can just go and pick it up and bring it to you.

MIKE SHEBANEK: That's going to be weird, isn't it?

SAQIB SHAIKH: It's going to be so weird.

MIKE SHEBANEK: Like today, I would think, hey, where did I put my keys to my house or something? Where did I leave my keys? It's like, oh, it's over there on the nightstand next to the bed, you know, 13 feet to your left. And in the future, I'll just say, go get my keys. Like, that's crazy. That's amazing.

SAQIB SHAIKH: I do think it's going to come in a way that we're not even going to notice it. It's just the devices are going to become smarter and smarter. Integrate, you know, with the AI. And one day they'll be like, wow, there's this new assistive tech device which does such and such. And so, yeah, I'm very excited about robotics and humanoid robots today. But yeah, I think they'll just be consumer devices for, like you say, very affordable tomorrow.

MIKE SHEBANEK: Yeah. So are you thinking that they're, because you mentioned it, they'd be gadgets. Are you thinking that there would be like a general purpose robot that could sort of, you know, have fingers and limbs and walk around? And pick things up and lift things and open things? And they would just do all those things? Like consolidated? Like think of, you know, on a smartphone, we used to have apps, you know, everything's on like one device, right? And it's all consolidated, if you will. Or do you think we'll have lots and lots of really task-specific robots? Because I think of it as we're going to have one humanoid-style robot that just does things. And I'm trying to imagine it really will be a world of like super specific robots. To do just one or two or three tasks.

SAQIB SHAIKH: I think it's going to be task-specific. But that's, you know, it's a guess. Who knows what's going to happen? I think there'll be lots of these humanoid do-everything robots that try to be like people. But I do believe that that's going to be what captures the popular imagination because it's really, really, really cool.

MIKE SHEBANEK: Yeah.

SAQIB SHAIKH: But I also think that, you know, a coffee maker is by, you know, if you go back some decades, people would look at a coffee maker today. And think, wow, that's kind of got robotics in it. So, yeah, I think we will just end up with smarter and smarter machines for each task. And that allows them to be more specialized and have better experiences. But you never know. There might be a do-everything one as well.

MIKE SHEBANEK: Yeah. Yeah. Who knows? We'll see. Do you think that it's going to take a new skill set? Like today we think about, you know, orientation, mobility, using a cane, learning Braille, learning a structure. Like there's a certain set of skills if you want to be employed or if you want to just be super productive online or things like that that we need to get. But in this world of AI, we can talk to it. We can give it pictures. We can ask about videos. We can ask it to generate things that don't exist like create a photo for me, make a movie, make music. Right? We haven't talked about music. But that's a whole other world of amazing AI application that blows my mind because some of that music is amazing. Do you think they're going to need to like have a new skill set for people who are blind and low vision to say like this is the new way to utilize all these tools and services? Or do you think that will just be sort of the machines will adapt to us?

SAQIB SHAIKH: I think a bit of both. So this is a bigger societal question. So if you take this away from accessibility and disability for a moment, then there is a question about if you want to know information. It used to be you look things up in books. Then you just did a web search. And now you just ask a chat bot. And what are the core skills that kids need to have? So, you know, schools still teach handwriting. And you could argue that for some amount of time that's going to continue because it's important in general. I think learning to read Braille is really, really important. Even after these AIs can just find information. And talk to you. I also think, you know, being productive might still need keystrokes and so forth. So, yeah, I think it very much applies to accessibility and the blindness community. But it's also this bigger question of when the AI could do everything for you. What are the fundamental skills as humans we should still know how to do manually? And yeah, a lot of good people thinking about that.

MIKE SHEBANEK: So, if you could jump forward in, say, three to five years. I mean, in technology, that's forever, right? Like that's who knows what the world like on three to five. A year or two, we can kind of guess. But three to five is like, ooh, it could be anything. If you could jump ahead three to five years and imagine an AI or AI technology, a robot, a person, whatever it is, and bring it forward to today, what do you think? What would you want the most? And then secondly, what would you want? What do I think would have the most impact today if I could kind of pull it forward? Or are we there now? Like nothing and just let life organically roll along and we'll be fine.

SAQIB SHAIKH: No way. We're not there at all.

MIKE SHEBANEK: I didn't think so. But I wanted to hear your thoughts on that.

SAQIB SHAIKH: It's crazy. Like I spend my days looking at all this new technology, the new AI, and it's capable of so much. But then, as I said, it's about helping people in their lives in the way they need. And while there are, you know, as someone who's blind myself, you know, there are people who do all sorts of things in life, you know, all the way to climbing mountains or whatever. But there are also those, you know, little road bumps in life, which I think can be smoothed out through technology. And there's still a significant chunk of those where, as someone who's blind, it takes more steps or more expertise. And if we can smooth out those hurdles, then you're going to enable and lift up even more people. And so, you know, I sit down and I can think of dozens of such examples. And to me, each of those is an opportunity. So, as I said, I think of my job as talking to the blindness community, talking to the scientists. And we still see so many of those challenges which can't be solved today. But if I try and imagine a future where the AI understands the world around me, understands me and my preferences and needs, and with robotics maybe can act on the environment around me, or act on the digital world when I'm at my computer, then, yeah, I just think we're going to get more and more power. And it's going to be so exciting. I can't wait to see what we're going to be able to enable and help people to do in the future as well.

MIKE SHEBANEK: Yeah, it is really quite incredible to imagine. Maybe harder for us since we're so rooted in the old world and the old technology. For young people coming into this field or getting started in accessibility and AI, their imaginations are going to run like crazy and it's going to be amazing to experience. I think one of the things that I love the most about the application of AI today and certainly in the future is that it's bridging the digital and the physical worlds. Like things like seeing AI where you can point to camera and say, 'What's in front of me?' Or, you know, be my eyes and be able to sort of ask, 'Oh, or be my AI in this case, and ask information, have questions, but also like actually move through the real physical space and have the AI apply to that because not everything is digital, right?' And that's what's exciting to me about Microsoft's, you know, seeing AI and Be My Eyes, and Aira, and some of these other types of applications in the here and now.

SAQIB SHAIKH: Yeah, totally. Like it's really about helping you. Yeah. In the moment, wherever you are, whatever you're doing. And yes, sometimes that's going to be making digital information accessible so that you can write your document or book travel or do something online. And then, you know, when you stand up, you want the AI to keep helping you. So, and then, you know, so many other things like we're thinking about audio description. So when you are in your living room in front of a screen where maybe the media was not made accessible, what does that mean? Or if you're going to an event or venue where there's not accessibility, what does that mean? So it's really, you know, the end, the whole experience of life. And that's why it's so much fun. It's like there's a really powerful tool getting so much better every few months. But there's so many interesting problems that we can make the lived experience even smoother. Yeah.

MIKE SHEBANEK: So much better. Right. And we're blocked in so many ways with the current technology. And AI really just unlocks all of those opportunities to address these things in ways that we couldn't have dreamt of before it. It's pretty absolutely amazing. Well, for those that are in our audience and thought, you know, maybe it's all been done. There's nothing left for me to do. It's all been solved. And we're talking about this like it's just a fait accompli, like it's already been done and now it's easy. I think the two of us would agree. There's a lot to be done. But now there's so many more tools and ways to get it done and to solve these problems and to make the world a better place. More inclusive, certainly, and more enabling for people with vision loss. So thank you, Saqib. It's always a joy to talk with you. You always have amazing and interesting perspectives and insights. And I could talk with you for hours, but I think we'll call it a close for the day. So thank you for participating in this and thank you to our audience for joining us. We really appreciate it.

SAQIB SHAIKH: Thank you so much. It's been a lot of fun.

[MUSIC PLAYING]