Why the Future of Intelligence Needs a Body and a Heart | Rajnish Sharma | TEDxZizhu Park

[00:00:03]Good afternoon everyone.

[00:00:05]Namaste.

[00:00:07]I wanted to start with a simple question.

[00:00:10]What is intelligence for?

[00:00:13]For most of the humanity, the answer has been very simple.

[00:00:17]Intelligence helps us survive.

[00:00:20]Intelligence helps us solve problems.

[00:00:23]In fact, intelligence helps us take on work which is difficult, and dangerous, and sometimes impossible to be done alone.

[00:00:35]That's why the story I want to tell you starts 500 m below the surface of the ground, below the earth.

[00:00:45]I'm going to transport you all into somewhere in 1960s.

[00:00:50]You are in somewhere in central India.

[00:00:54]This is an underground coal mine.

[00:00:56]What you see here is hundreds of miners who would every day take this mining elevator and go down the shaft.

[00:01:06]Every day spend 8 hours below the earth to extract coal.

[00:01:12]All the miners used to call coal black diamond because coal used to power homes, it used to power factories. In fact, coal powers the entire economy. They were extremely extremely proud of what they did.

[00:01:29]In fact, my father started his career as a mining engineer.

[00:01:35]You see him here, a handsome young man just out of his engineering college.

[00:01:41]And he used to work in one of those underground coal mines.

[00:01:44]Very difficult work.

[00:01:46]There'll be him and his team which will go down every day hundreds of meters below the ground, and for 8 hours 365 days year, worked there.

[00:01:59]As you can see, under the ground, you have trolleys, you have tracks, and people would manually use the tools to take the coal out and haul them above the ground so that the industries can be powered.

[00:02:13]I was a kid then.

[00:02:15]I didn't fully understand what he did.

[00:02:18]But, I understood what my family felt.

[00:02:22]Why?

[00:02:23]Because, I remember my grandmother every day would pray when he left for work.

[00:02:31]Because, when somebody disappears so far below the surface of the earth, there's a uncertainty.

[00:02:37]There's a risk.

[00:02:38]And that is what it is.

[00:02:41]But, it was a very, very difficult work.

[00:02:44]It was a dangerous work, but it had to be done.

[00:02:47]And did that generation pretty much did it.

[00:02:51]In fact, you know, I didn't realize it then, but that image stayed with me forever. In fact, I still recollect I was a kid. When he used to come back, I used to run to him and he'll say, "No, no, go away.

[00:03:03]Let me take a bath first." I could still, you know, I have a a sense of a smell, the coal dust on his clothes, that still stays with me.

[00:03:12]But then, I didn't realize that did that thing stayed with me even today. But, if you look at it, that is the history of the humanity, right? We find a burden, we find a problem, and then we really work hard to solve it. And that's what it is all about, right? We We there's a problem, we all work towards it and solve it. Let's move further.

[00:03:34]25 years later, 1993, I was an engineering student, and I was working in on my final year dissertation. What was my topic? Expert systems. In fact, the photo you see on the right is a photo of the project report I submitted.

[00:03:51]Back then, artificial intelligence were not a buzzword.

[00:03:56]And what fascinated me and my team was, you know, can machines really think?

[00:04:01]Can they behave like expert?

[00:04:04]Can they make decisions?

[00:04:06]Can they replicate human judgment?

[00:04:09]It was a very futuristic question at that point in time.

[00:04:12]Today, it feels very quaint.

[00:04:15]It feels very naive.

[00:04:17]Because today, we are not really thinking about can machines think. We are talking about can machines act?

[00:04:28]And people, you know, normally ask me if AI has transformed so much, why is everyone talking about embodied AI?

[00:04:38]We are talking about embodied AI because the digital AI, as we all know it, it can pretty much analyze data, it can analyze patterns, it can even reason, but it cannot interact with the physical world around it. That's a fact. And that is what is changing. Now we are teaching machines how to act, how to perceive, how to really move around, and sense the physical world around them. That is what is embodied AI all about.

[00:05:10]And we see this in multiple ways, right?

[00:05:13]I think most of us know we see these robotics arms in the factories, we see these EVs all around us, we know about drones, we also know about humanoid robots. This is what embodied AI is all about. In fact, the evolution of robotics pretty much resembles the evolution of intelligence itself.

[00:05:36]Because at first, we had machines that could simply follow human instructions, fixed instructions, right?

[00:05:43]And then came machines who would understand the surroundings and navigate the surroundings around that.

[00:05:50]Thereafter, we had machines who could learn from human beings. But today, we are entering an era where vision, reasoning, language, and motion, they all are coming into a single system. And when this happens, the intelligence begins to move out. It begins to move out from the screen into the real physical world. And when that happens, there is the potential of artificial intelligence to as profoundly impact our physical world as the digital AI did our information world. You know, kind of wondering about this question, will silicon rival carbon someday?

[00:06:31]What it really means in, you know, layman's language is, will there be a day where machines will be as capable, if not more capable, than human being?

[00:06:40]And with endo with the advent with the arrival of embodied AI, I think we are getting there because we are now having machines which can pretty much perceive, which can pretty much understand, and they are trying to learn to navigate into the world around them. And that's why many of us, like me, believe embodied AI, all this that you see here, it's going to be the most potent, most powerful productivity force this generation has ever seen. But amidst all this excitement, you know, we all miss a point. This thing is not about robot, robotics. This thing, I believe, is about trust. Why do I say that? Let me explain this. My kid, my son, he's 21 years now by the way. When he was a kid, his most favorite transfer, you know, toy, was a Transformer robot.

[00:07:37]I think you all know them. They are powerful, they are very capable, and they have this ability to transform into something else when needed. He used to simply love it. In fact, he used to talk to it. He used to take the toy with him, the Transformer robot, everywhere he went. In fact, he used to sleep with it.

[00:07:57]So much so that he wanted to become a robot, a Transformer robot, when he grows up.

[00:08:03]In fact, this is a pic of him in Shanghai in 2009, most probably, where he actually became a Transformer robot.

[00:08:08]He tried very hard.

[00:08:10]Now, I used to wonder, what is it in that harmless toy that kids place so much of trust in that?

[00:08:20]They placed so much trust in them because they found them capable, they found them reliable, and they found them very helpful.

[00:08:28]They did not, you know, look at them just because they were sophisticated, but these are the qualities they looked at it. And that thing stayed with me for a very, very long time. In fact, very recently, we had COVID and I think during that time, first time ever, we saw the trust that the humanity placed on the physical machines.

[00:08:51]In fact, living in Shanghai during that time, some of us would have realized and we were fascinated by the power of machines during that time. We actually saw machines disinfecting public places. We had service robots delivering uh food to the, you know, different uh hotels.

[00:09:10]We also had uh robots assisting health care workers.

[00:09:14]We all trusted them.

[00:09:16]Not because they were sophisticated.

[00:09:20]We trusted them because they were reliable. They showed up when needed and they reduced the risk. And that for me is the biggest lesson that AI can take, that the trust comes before technology.

[00:09:34]Trust comes before adoption. And trust is what makes the technology being adopted. I think that's a That's really is the lesson that, you know, we can take forward.

[00:09:46]In fact, we are seeing the signs of it in embodied AI, right?

[00:09:51]I've spent almost close to 30 years in the industry in the technology industry and with the privilege of leading one of the very large and successful technology company in China, I see myself witnessing another inflection point.

[00:10:05]And that inflection point is all about digital or embodied AI, which is digital AI becoming physical. We see the signs of it, right? We see robots now in warehouses, logistic centers, in hospitals. We see them see them in elder care facilities. So much so, if you recollect, we actually recently saw them running marathons.

[00:10:28]And they In fact, they were better than the human beings. We see them performing and the gala, the New Year gala. I think the fantastic performance that was done, we all remember it and we are going to remember it for a long, long time.

[00:10:42]All these moments capture our imaginations, but I think these are just the signals. These are the signals that the intelligence is stepping out of the screens and getting into the physical world. I think that is the that is the most important thing. And when that happens, you know, people have questions in their mind.

[00:11:03]They feel that, you know, what are the different possibilities? What is it that these things can do for us? And we think, you know, there are many unanswered questions. What will happen if when robots get into the factory? How will they behave? What will be the contracts between the human beings and the robots? How will they interact with the existing systems? There are many, many unanswered questions. In fact, if you look at the last two or three years marathons, we had so So videos floating around where the robots are falling and then they get up again and you know so so many things. So there are a lot of hiccups. But we are getting there.

[00:11:41]And one of the lessons that I've learned being in this industry for so long is that nothing comes easy. We all learn through iteration. And machines will be no different. If you recollect, humans don't learn uh They are not perfect, right? You don't get the breakthrough just waiting for the uh perfection.

[00:12:01]Human beings learn through experiences.

[00:12:03]Our kids children's learn through experiences. And therefore machine is going to do the same. And that's what will happen with uh with the robotics and machines also because every time they do this, they will learn. And the future will belong to them who will deploy, learn, improve, and deploy again.

[00:12:25]What I really mean by this is whenever machines interact, they create more data. As you have more data, you have better intelligence. Improved intelligence will create better experiences. And with better experiences, once you have responsible AI and empathy built into those experiences, you are going to create more trust and trust will drive more adoption. That is how the things will go on. In fact, that leads to the virtuous flywheel. Right? What is that virtual flywheel? For any of these artificial intelligence implementations you start with empathy.

[00:13:02]You If you have more empathy, you'll have more trust. If there's a trust, there's more adoption. That leads to more data being generated and therefore there's an improvement and finally it reinforces trust. I think the lesson to be learned here is it starts with empathy.

[00:13:18]And then there has to be trust built in.

[00:13:20]And this is really how the embodied AI is going to scale. It's not going to scale only by algorithms. You need to have a human-centered intent and a human-centered design to scale embodied AI, and that's what will happen going forward. But, you know, with every technology revolution, there is an anxiety.

[00:13:45]People ask, "What about jobs?"

[00:13:47]I think that's a very, very fair question.

[00:13:50]But, humanity and the history of humanity tells us that whenever there is a technology revolution, it never eliminates human value.

[00:14:00]It redefines it.

[00:14:02]Let's go back. Electricity did not eliminate work. Steam engines did not eliminate work.

[00:14:08]Computers did not eliminate work. What they really did is they redefined the nature of the work.

[00:14:15]I think robotics is going to go the same way.

[00:14:20]It's It's job is to first do the work which is dull, which is dangerous, and which is hazardous. The kind of work my father's generation had to do because there was no alternative.

[00:14:33]The machines are not supposed to replace human beings. They are going to uplift them.

[00:14:38]The job is not replacement. The job really is amplification.

[00:14:43]It is all about extending the capability of human beings, and that's what it's going to be.

[00:14:52]2 years back, my father passed away.

[00:14:56]In fact, this is uh my last picture with him, just 3 months before he passed away.

[00:15:02]And I wanted to do this talk to pay my tributes to him.

[00:15:09]If I think about it, he belonged to a generation that pretty much powered the entire industrial revolution through their hard work, through their courage, and through their sacrifice.

[00:15:21]We all, today stand on the shoulders of that generation. They did that hard work, that's why we are here.

[00:15:29]And no amount of gratitude is good enough to pay our tributes to that generation.

[00:15:35]And when I think about the time that he spent in 1960s working and trying to get, you know, energy out of the earth and what we are doing today, I wonder because he spent all his time getting energy out of the earth, here we are spending our time in getting intelligence into the machines.

[00:15:56]Now, on the surface of it, this looks very, very different. But the more I think about it, they are very, very connected. They are connected because they all are connected by the common purpose. And that common purpose is to do human good. The common purpose is to make life safer, better, and really expand what humanity is capable of achieving. Amplifying human potential. I think that's what this is all about. We have seen this before. We had tele- telescope that amplified our sight. We had ships that amplified our rea- reach. We had uh computers that amplified our mind. We have internet that amplified our knowledge. And I'm sure robots, one day, are going to amplify our ability to do work. We are seeing the signs of it. We see them in elder care facilities. We are seeing them doing very precise work in factories. And we are seeing them in the mines, something which is very, very close to my heart. And that was this uh robotics going to do. And that it was the signs of things to come. And that's a good news for all of us. And therefore, what I believe is one day, if we are successful, robotics is going to be successful only when if they are going to work towards not replacing us, but not replacing humans, but making us better humans. And that is why I believe that the future of intelligence needs more than a body.

[00:17:29]What it needs is a heart.

[00:17:32]Thank you.

[00:17:33]>> [applause]