AI is polarizing — you see it referenced in pop culture as a savior (Star Trek) or a threat (Terminator). Or similarly, in the media — with positive news about drug discovery and health implications, and disturbing news regarding facial recognition and the potential of autonomous drones. Should we remain optimistic about a future increasingly reliant on AI? Why?
When I tell people that I am a robotics researcher, I tend to get one of two reactions. Some people get nervous. They make jokes about Skynet, and ask when the robots will take over their jobs. They worry about autonomous vehicles because they don’t trust computers to avoid crashes as well as they could themselves.
The other group gets excited. This is the group that buys Roombas to vacuum their living room, and dreams of the day when their car will drive them home from work. They can’t wait to hear about the latest research, and discuss all the ways that robotics and artificial intelligence will change our lives and our world for the better.
As the Head of the Computer Science and Artificial Intelligence Laboratory at MIT, I’m sure you won’t be surprised to hear that I’m in the latter group. But I also know it’s an important part of my job to understand the fears of that first group, to listen, to ask questions, and to give them some perspective on why I see things differently. That starts with understanding that AI is just a tool. It’s an incredibly powerful one; but- — like other tools — it isn’t inherently good or bad. It is what we choose to do with it. And I believe we can do some truly incredible things.
On a global scale, AI will help us generate better insights into addressing some of our biggest challenges: understanding climate change by collecting and analyzing data from vast wireless sensor networks that monitor the oceans, the greenhouse climate, and the plant condition; improving governance by data-driven decision making; eliminating hunger by monitoring, matching and re-routing supply and demand, and predicting and responding to natural disasters using cyber-physical sensors. It will help us democratize education through MOOC offerings that are adaptive to student progress, and ensure that every child gets access to the skills needed to get a good job and build a great life. It may even help those kids turn their childhood dreams into reality, as Iron Man stops being a comic book character and becomes a technological possibility.
On an individual level, AI will offer opportunities to make our lives safer, more convenient, and more satisfying. That means automated cars that can drive us to and from work, or prevent life-threatening accidents when our teenagers are at the wheel. It means customized healthcare, built using knowledge gleaned from enormous amounts of data. And counter to common knowledge, it means more satisfying jobs, not less, as the productivity gains from AI and robotics free us up from monotonous tasks and let us focus on the creative, social, and high-end tasks that computers are incapable of.
All these things — and so much more! — become possible when we direct the power of computing to fix the things that humans haven’t been able to fix on our own.
I asked this question to Dan Huttenlocher and am also interested to hear your thoughts — President Reif [of MIT], in an early note about the new College of Computing, mentioned that its students will be “navigating an algorithmic future.” That is a provocative concept. How do you envision our algorithmic future? Is it one in which algorithms recede into the background, helping us in invisible, but profound ways? Or one in which the algorithms supplant facets of the physical world? A mix? Or something else entirely?
Today telepresence enables students to meet with virtual tutors, and doctors to treat patients thousands of miles away. Robots help with packing on factory floors. Networked sensors enable monitoring of facilities, and 3D printing creates customized goods. We’re surrounded by a world of possibilities — possibilities that will only get larger as we start to imagine what we can do with advances in artificial intelligence and robotics.
Picture a world where routine tasks are taken off your plate. Fresh produce just shows up on your doorstep, delivered by drones. Garbage bins take themselves out, and smart infrastructure systems support automated pick-up. AI assistants — whether embodied or not — act as our guardian angels, providing advice to ensure that we maximize and optimize our lives to live well and work effectively.
How will we live in the future? How will we work? How will we get there? How will we get paid? What will we do for leisure? How will we ensure that everybody’s life will be better? Many are thinking about these questions.
Technology has the potential to support us in how we live and how we work, from small things — like making riding to work easy and safe — to the much bigger things. As a society, we have a lot of big challenges to solve: inequity, poverty, hunger, affordable housing, health care, education, climate change, fake news, and everything in between.
Technology in general — and AI in particular — can be an incredible vector for positive change as we work together to figure these things out. It can bring us together and give us ways to navigate the truth in a world inundated with alternative facts. It can take care of the routine stuff, and that gives us more time to focus on solving the big challenges that require creativity, collaboration, and strategy.
I know this sounds optimistic. AI is not going to solve all of our problems any more than it is going to destroy the world. Today AI can crunch numbers, and remember, and make predictions better than people can, but AI can’t learn, communicate, nor understand the world like we do. It doesn’t have common sense. This doesn’t mean AI isn’t a powerful tool. Only that it isn’t either/or: humans or machines, consider instead humans and machines. Working together with AI systems, humans can augment and amplify many aspects of work and life.
How has your perspective on AI, especially as it pertains to robotics, evolved since your time as a doctoral student at Cornell?
20 years ago computation was a task reserved for the few. Computers were large and expensive and hard to access. A person needed a level of expertise to know what to do with them. This began to change with the advent of home computers. Suddenly, every office and classroom had a computer sitting in it.
Still, most people could only use these computers for very limited number of tasks: type up documents, browse the Internet, maybe play a few games. But all that changed with the smartphone.
The smartphone democratized computation. Suddenly, there was an application for every task — an application that anyone with any level of knowledge could use. Computation became the new normal — so normal, in fact, that we stopped noticing how central it was to almost every part of our lives.
I believe we are about to experience this type of change again. But this time, it will be AI, ML, and robotics that transforms how we live and work. Machines will help us deliver on all types of cognitive and physical tasks, and just as with computing, the benefits will be shared by people of all ages and education levels.
Which sector(s) do you see being most affected by autonomous robotics and systems in the short term (2–5 years) and the long term (10+ years)?
Transportation, logistics, and warehousing are great examples for 2–5 years impact. It’s much easier to move a robot through the world than it is to build a robot that can interact with it.
Over the last few years, significant advances in algorithms and hardware have made it possible for us to dream about the ability to move people and goods in a much safer, more convenient way. This is all possible because of rapid advances in technology.
Today, we can do so many simultaneous computations, crunch so much data, and run algorithms in real-time. We’ve built better hardware components including sensors and actuators. We’ve advanced our understanding of how to make maps, localize information, and plan routes. These technologies have taken us to a point in time where we can realistically discuss the idea of autonomy on the roads.
But not full autonomy. Not yet — this is for a longer timeframe. The perception system of a vehicle doesn’t have the same quality and effectiveness as the human eye. To be clear, there are some things that machines can do better than people, like estimate accurately how quickly another vehicle is moving.
But robots don’t share our recognition capabilities. How could they? We spend our whole lives learning how to observe the world and make sense of it. Machines require algorithms to do this, and data — lots and lots and lots of data, annotated to tell them what it all means. To make level 5 autonomy possible, we have to develop new algorithms that help them learn from far fewer examples in an unsupervised way, without constant human intervention.
Looking further out, robots will revolutionize production. Today, robots have enabled mass fabrication. In the future, they will enable a world of customized production where making will be easy, intuitive, and pervasive, much like printing is today. Eventually, robots won’t just be producing items of any shape and size; they will become them.
Imagine the ability to program matter to change shape or even become invisible. It may sound like the stuff of science fiction, but I believe that eventually, we will be able to program robots to change a wide range of mechanical properties — for example changing optical properties to become invisible, changing acoustic properties to move in “stealth” mode or to enable easy localization, or visco-elastic properties to become vibration-absorbent or to allow high jumps.
There is also great potential for doing more for healthcare. For example, robots can measure vital signs in the case of highly infectious patients. MIT’s e-vent project is an open source project for an easy to fabricate low-cost robotic device that turns AmbuBags, which are broadly available, into emergency ventilators — which are in short supply at the moment.
These are 3 example sectors — many other industries will be greatly impacted by this technology.
What innovations in robotic hardware excite you most? What makes them transformative?
I am particularly excited by the transformative potential of soft robotics. While the past 60 years have defined the field of industrial robots and empowered hard-bodied robots to execute complex assembly tasks in constrained industrial settings, I believe the next 60 years will be ushering in robots in human-centric environments and our time with robots helping people with physical tasks.
While the industrial robots of the past 60 years have mostly been inspired by the human form — the next stage will be soft robots inspired by the animal kingdom with its form diversity and by our own built environments, with broader applications potential. With the development of soft materials, machines and materials are coming closer together with machines becoming compliant and fluid-like materials and materials becoming more intelligent and this raises an interesting question: what is a robot?
Traditionally we have considered industrial manipulators and robots on wheels but what about robots made our of food or paper or ice, or giving everyday objects the ability to move and compute so they become intelligent and autonomous?
Will AI, like calculus or geometry, become a prerequisite for every engineering program?
I believe that computational thinking should be part of literacy in the 21st century.
Finally, we’ve seen our investments coalesce into three general groups, those companies whose core technology will help solve climate challenges; those that will create new human health solutions; and those that will usher in computing & systems of tomorrow. Briefly, where will we see the greatest impacts of AI in each of these groups?
AI is instrumental in addressing climate change. New innovations can slow the impacts of warming, whether through stopping deforestation, preserving biodiversity, or ensuring there’s enough food to go around and food is not wasted. But there’s also an opportunity to address some of the underlying causes. Our planet is a feedback system. Reducing carbon emissions is not enough to reverse climate change. We need to cool the planet. Instead of aiming to limit global warming temperature increase by 1 degree Celsius, we have to think about what we’d need to do to reduce temperatures by 1 degree Celsius. The solution will be at the intersection of Robotics, AI, Materials, Environmental Studies, and more.
In healthcare, AI is enabling new approaches to better diagnose, treat, and monitor disease, toward an era of fully individualized healthcare, where medicines will be synthesized for each person as they need it and treatments will be customized.
AI, ML, and robotics will continue to transform how we live learn and work. Machines will help us deliver on all types of cognitive and physical tasks, and just as with computing, the benefits will be shared by people of all ages and education levels.