Dr Rand Hindi is an entrepreneur and data scientist. He is the founder and CEO of Snips. This is the full transcript of his TEDx talk titled ‘How Artificial Intelligence Will Make Technology Disappear’ at TEDxÉcolePolytechnique conference.
Dr Rand Hindi – Founder and CEO of Snips
So, just before I start, I’d like to let you know that I actually do have a PhD, so everything I say must be true. OK.
About a month ago, I went on holiday in Costa Rica with my girlfriend. And it’s a really awesome place, it’s basically jungle and beaches. It’s jungle and beaches, and you spend your entire day without being connected. There is no Wi-Fi, there is no 3G. And I don’t know when was the last time you felt this kind of peace in a big city. We cannot forget what nature feels like.
The problem is that, as soon as we got back to the hotel, we got connected again. There was Wi-Fi. And as soon as we got connected, we started getting interrupted. All of a sudden, all these push notifications from our phones would start interrupting us. It was interrupting us when we were having dinner. It was interrupting us when we were having our showers, actually going and pressing with your wet finger.
But it was also interrupting us when we were making love. And this, for me, was kind of a big problem because, not only was I interrupted, I also kept on reaching out for my phone, hoping that perhaps, there was a notification. And I wasn’t doing this consciously. It was exactly like Pavlov’s dog, but my bell was my ringtone, and my sugar was that one notification a day you might get that makes you happy. Perhaps is a message from someone you secretly love, perhaps is Justin Bieber following you on Twitter.
As it turns out, we got so addicted to technology, that 9 out of 10 people today experience something called ‘phantom vibrations’. This is when you have your phone in your pocket, and you thought it vibrated, you pull it out, there’s nothing. If you’re laughing, it means it probably happened to you. Nine out of ten people. But think about it. We got so conditioned by technology without knowing it, that we’re now hallucinating 90% of the time.
So, how did we actually get here? I’ve got a great, super low-tech thing here. If you look back at the history of connected devices. So here you’ve got time, and here you’ve got whatever-that-will-be. Back in 1990, you didn’t have any connected devices. This was what you called “the unplugged era”. 1995, you now have something called “the Internet”. And this was awesome, because all of a sudden, all these computers got connected. So you had one connected device.
But with this first connected device, came something called “e-mail”. E-mail also came with the first notification which was the “You’ve got mail”. Sometimes, it was a friend, sometimes, it was a picture of a dog, sometimes it was spam. Most of the time, actually. So that’s the Internet era.
Ten years later, 2005, that’s what you call “the mobile era”. This is what we’re exiting right now. And this time it’s not one, it’s three devices which are connected. It’s your phone, it’s your computer, and it’s a tablet. But with three connected devices, you’d think that you’d actually have more value, but what really happens is these devices are not smart enough to figure out which one you’re currently using, so the default strategy has been to push everything to all devices. If you have a phone and a computer and someone calls you, it rings on both. If you answer it keeps ringing, because your computers are not aware of each other. This is a big problem because we’re now entering something called “the Internet of things” era, IoT.
The Internet of Things is when your fridge is connected, your watch is connected, your car is connected. And this time it’s not going to be 3, it’s going to be between 10 and 15 devices per person. There’s going to be more than 100 billion devices connected by 2025. A hundred billion.
Just try to imagine a little bit, what 15 devices that you own will feel like when all of them are requesting your attention, when they’re interrupting you all day long. And when the only way you have to actually stop the interruption is manually interacting with each of them. This is horrible, this is just horrible. It’s like the anti-Costa Rica.
And when you actually look at this, I’ll just plot this in here, you actually see that this curve is actually exponential, which means that if you keep going, let’s say 2035 or whatever, you might have maybe 1,000 devices. 1,000 devices. You’re going to get so many notifications you’re probably not even going to hear them. It’s just going to be background noise now. So, clearly, we don’t want that kind of future. This is a real problem, and I’m saying this because people don’t realize that just yet. They think that this is just how technology works. But this is not how it works.
There is something else happening right now which is called “Artificial Intelligence”. And I’m not talking about the robot kind that will kill us all, I’m talking about a specific type of AI called “Context Awareness”. Context Awareness is the idea that you can give your devices the ability to sense and react to the situation you’re in. So in this specific scenario, Context Awareness is the anti-friction, because context awareness is how you make your phone not vibrate when you’re in a meeting. It’s how you don’t have to actually click on an interrupter when you get in your house. The only thing is that AI takes a little bit longer to start, so it’s kind of taking off recently, but actually, when it does take off, it goes much, much faster.
So right now, for the next few years, we are going to get more and more friction. It’s going to be horrible. Whatever you think you’re experiencing now, it’s going to get much worse in the next few years. But at some point, when AI will become so capable that the value it brings will be more than the friction of all your devices, then, something really amazing is going to happen. The overall frictions, so the experience of technology, which has increased in the next few years, all of a sudden, will just revert. And because the AI grows super rapidly, it’ll go back to zero very fast.
There is a time in the future where technology will have no more friction, thanks to AI. This era is what you call “ubiquitous computing”. Ubiquitous computing is about being able to add a new device which doesn’t add friction; it actually adds value. And this is what the digital revolution is all about. It’s about using this artificial intelligence to make technology disappear in a way that you can just go about your day and not care about it anymore.
So is this really such a crazy idea? Well, I think it’s pretty cool. But when you go back a couple hundred years ago, there was another very intrusive technology. It was really bad at the time. It was called “electricity”. Electricity, in the beginning, was very difficult to produce. It was very expensive, it was unreliable, it would cut all the time. It was dangerous; you got electrocuted, your house would catch fire. This was the reality of electricity back in the days. People thought that oil lamps were actually safer. Who would ever want to put any electrical stuff in their walls and make everything go in flames?
As the technology matured, it became cheaper. It became more reliable, it became more accessible. And today you have it everywhere, you have it in your walls, you have it in your phones, you have it in your heart if you have a pacemaker. You don’t think about electricity anymore. It became ubiquitous, it’s all around us. And we’ve integrated it in our lives in such a way that we forgot about it, it disappeared from our consciousness. And this is exactly the same thing happening with computing.
Computing will disappear. And why am I excited about this? I’m excited because it’s not just a question of having your tea already prepared in the morning because your bed is connected to your coffee machine or whatever. It’s not just about your car driving itself, and actually knowing where you’re going. It’s not about that. It’s actually about having the freedom to actually enjoy the world in a way that feels unplugged, but yet, has all the value of technology.
Now, just to prove to you that this is not science fiction, I’m going to explain how we actually built it. Context Awareness: the first important thing is being aware of the people you interact with. This here is an example of what my e-mail connections look like. The larger the bubble, the more e-mails we’re actually exchanging. So the big thing in the middle here is one of my co-founders. He’s been sending me a little bit too many e-mails after we did this a few times. But what’s interesting is that, if you start mining those social relationships through e-mail, you can start not only figuring out what sort of relationship you have with people — professional, personal — you can also start predicting when you’re about to get in touch with them. And if you can predict who you’re going to be talking to very soon, perhaps you can automate a lot of these really user messages like On My Way. They’re small things, but they can be automated.
You can also start looking at things like your calendar. A calendar is a great way to have an idea of what you plan doing during the day. But nobody really writes through calendar events in a very comprehensible way, because the context of the event is in your head. Now that your computer actually knows how you’re interacting in which context, it can figure out that specific meeting which says “Meeting with Michael” is referring to the same Michael that you exchange e-mails with in that context, and therefore, it can prepare the meeting for you with that person.
But your calendar is still not your life. Here is probably the most interesting data that you can ever get. It’s location data. This is what your phone sees when you’re moving around the city during the day. But location data is very noisy so what you need to do is analyze it, so you need to be able to differentiate between moments when you’re in a space, in a place, so knowing that you are in a restaurant, or at the gym, is a very, very important contextual feature to understand what you’re currently doing. You also need to know how you actually go between places knowing that you like to bike on your way home is important because this is how you make sure that there is a bike whenever you need to take one to go home.
And if you do this over a long enough period of time, you can actually start figuring out people’s habits. You can start figuring out that they like to go to the gym at 5 pm after going to the office. And that they like to walk there. And that when they get there, they like to use this particular machine. So you can start thinking about a place where your gym would be smart enough to actually schedule the machines according to all the people that will be coming in, without you actually having to ask for it. And this is what we talk about when we talk about ubiquitous computing.
Now if you look at the environment itself; take this location trace. And now take it for 100,000 people. This is what we did in Paris. We analyzed the location data for 100,000 people to see how they actually flow in the public transport system. And this is beautiful, I mean, you can see the heartbeat of the city. You can see people going home, going to work. This data is more accurate than any measurement ever made.
And now that you have this, not only can you help people find a place in the train home so that they can be comfortable, you can help train operators actually figure out train capacity in a way that is optimal. You can do the same for car accidents. You can literally predict how likely you are to have a car accident based on when and where you’re driving. And this is important, because if you have your kids in the back of your car, or if your self-driving car is driving your kids to school, you want it to be safe. So you need this context about the risks surrounding it.
You can even do it for predicting queues at a post office. Why would you go to the post office and queue for half an hour if you can just wait ten more minutes and have no one there? Why would the post office actually tell you to come between 8 and 12, instead of actually predicting when is the right time for you to come? So, you see, Context Awareness is not just about figuring out what you’re doing, it’s about putting all these different layers together, it’s about putting you, and your environment, and the people you interact with in the same system. And what you get in the end is a very contextualized timeline of what you’ve been doing throughout the day with a prediction of what your intentions are.
And now that you can predict your intentions, you can give that intelligence to all your devices, so that they can start working together and decide what is the optimal way of interacting with you. Sometimes, it will still ask you for a confirmation, because it’s not going to take a shot unless it’s sure. But at least it’s not going to be 1,000 of them doing it at the same time.
I’m going to tell you a little secret. Everything I’ve just shown you we just built. We’ve been working secretly for over a year now, on completely reinventing the way that we’re using our devices. And we started with the smartphone. We’ve created essentially a radically new interface that leverages this artificial intelligence to anticipate everything you’re trying to do with your phone, so you barely have to use it. You can sign up on this link, and I’ll try to give it to you.
But what’s important here is that we cannot do this alone. Because we cannot decide how the interface for all of us will look like. We need actually your help, because without your help, we will not be able to figure out this artificial intelligence that will eventually go into all of those devices. And without this artificial intelligence, the future will be more technological slavery. And we don’t want that. What we want is a future where technology has disappeared. And where we’ll finally have the freedom to spend our lives enjoying time with the people we actually love.